Daxos Capital · Diligence Board

Plug and Play University Pitch Day 2026

All 36 companies from the NYU, Rochester, Rutgers & Stevens Alumni Pitch Session, pitching 2026-07-29. Every founder claim, funding figure and customer name was checked against sources outside the company. The top-rated names were then handed to a second set of agents whose only job was to refute the rating — that pass moved seven companies down, three of them by more than two points.

36
Reviewed
1
Funding figure verified
11
University tags wrong
7
Cut after refute pass
7.25
Top rating

How to read this

A rating is a confidence statement, not a verdict on the business. Verified elite pedigree and independently corroborated traction score high; claims that exist only on the company's own website score low no matter how specific they sound.

That rule cost the original top three their places. The lesson from the refute pass, in one line: a claim appearing only on the company's own site is unverified by definition.

Nothing here clears the elite-pedigree bar. No Palantir, SpaceX, Anduril, Tesla, Anthropic, OpenAI, Citadel, Jane Street or Stripe alumnus appears anywhere in the cohort with a verifiable senior role.

Worth a look

Verified something real. Still needs a founder call. · 3 companies
#CompanyRatingRaised (verified)Deck saysWhat it does
20Adomate Comm7.25€1.4M seed$1.6MAI data platform for ad creative at scale
33Aridica Corporation Deep7.00Accelerator only (~$10K on record)$2.7MPush-button immune-cell isolation in 30 minutes
36Parallel Carbon Deep7.00$3.6M seed$3.8MElectrochemical carbon capture co-producing cheap hydrogen

Adomate

adomate.com ↗
7.25/10
€1.4M seedU Rochesterdeck said $1.6MUniversity tag does not match

Ghent-based, founded 2024 by two ex-ML6 applied-ML engineers. Lets performance marketers scale creative output to meet algorithmic demand for volume and diversity.

What checks out

  • Funding matches public record: €1.4M seed, named investors (Angelwise, RDY, Seeder Fund)
  • Named brand customers: Mazda, Loop Earplugs, Neuhaus, Gentlee, Thrive
  • Three named-and-titled customer testimonials — claims are customer-attributed
  • Meta Verified partner

What doesn't

  • No elite-list pedigree — ML6 is a credible EU consultancy, not hard-culture
  • U Rochester tag matches nothing found; company is Belgian
  • "50+ customers" and "5-10x more creatives" self-reported

Aridica Corporation

aridica.com ↗
7.00/10was 7.50 before the refute pass
Accelerator only (~$10K on record)U Rochesterdeck said $2.7MDeck raise contradictedProduct not shipping

Replaces a 50-year-old manual 3-4 hour PBMC isolation workflow. CEO Palani Palaniappan came from GE Healthcare, Thermo Fisher and Epic Sciences; CTO Sergey Gershtein ran Abbott's i-STAT platform (200,000+ units installed) and is the named inventor on the patent filing. Advisors include Stan Lapidus (founder of Cytyc and Exact Sciences, 37 patents).

What checks out

  • Genuinely elite industry pedigree — the strongest team on the board
  • Advisor bench includes real diagnostics exits (Exact Sciences, Illumina ex-CFO)
  • Patent application filed: US20250269373A1, correct assignee
  • Selected 1 of 5 from 175+ for BioTools VANGUARD

What doesn't

  • $2.7M claim vs ~$10,000 recorded in PitchBook — only accelerator/non-dilutive money traceable
  • Flagship 30-min isolator is not shown shipping; only SKU is a cryotube adapter
  • No published performance data or comparison vs Ficoll
  • No SBIR/NIH grant despite being exactly the profile that wins them
  • Boulder CO, not Rochester

Parallel Carbon

parallelcarbon.com ↗
7.00/10was 7.25 before the refute pass
$3.6M seedRutgersdeck said $3.8MUniversity tag does not matchNo pilot data

UK company (London/Manchester) using alkaline-earth metal hydroxide looping — electrolytically generating acid and base to capture atmospheric CO2 while co-producing hydrogen. CTO Aranzazu Carmona Orbezo holds a Manchester PhD in chemistry and is co-inventor on the patent.

What checks out

  • Funding checks out: $3.6-3.8M across 3 rounds, led by Counteract
  • Patent AU2023212880A1 assigned correctly to the company
  • Real differentiator: process tolerates on/off operation, unlike most DAC
  • PhD electrochemist CTO with first-named patent

What doesn't

  • No pilot plant, tonnage figure, or third-party validation found
  • "Cheaper than fossil fuels" with zero published cost data
  • Founded 2021-22 and still only ~$3.8M — ask why no Series A
  • UK company, not Rutgers
  • Crowded field: Mission Zero, Sustaera, Verdox

Second tier

Real companies, thinner verification or weaker fit. · 6 companies
#CompanyRatingRaised (verified)Deck saysWhat it does
16Ferris AI Comm6.75$0$0Agentic platform for civil engineers and AEC firms
13Adomo.ai Comm6.50Unverified$650KEnterprise AI ops platform with human approval gates
23LiquidRails.ai Comm6.50Unverified$750KVertical B2B SaaS for beverage-alcohol DTC commerce
12Nexhelm AI Comm6.50$20K$20KAI execution layer for wealth-management firms
7Grego AI Deep6.00Undisclosed seed$2MAI smart-contract auditor for Web3
19TwinsAI Comm6.00Unverified$500KReal-time AI copilot inside live sales calls

Ferris AI

askferris.io ↗
6.75/10was 6.50 before the refute pass
$0RutgersPedigree unverified

Four agents covering document intelligence, calculation verification, submittal reviews and template generation, plus FDX, a data-management layer for AEC firms. Founder Daniel Calabro is a structural engineer turned founder.

What checks out

  • Domain-expert founder in a genuinely underserved vertical
  • FDX is the durable layer worth a second look
  • One named testimonial from a practicing structural engineer

What doesn't

  • No verified prior employer, title or degree for either founder
  • No Rutgers record found
  • Pre-revenue design-partner motion; "saves 3+ hours daily" unsubstantiated
  • Four people in a crowded 2025-26 AEC-AI cohort

Adomo.ai

adomo.ai ↗
6.50/10was 6.75 before the refute pass
UnverifiedNYUdeck said $650KStudio-runNo technical team named

Describe a multi-step process in plain English; the system plans it and executes across 30+ integrations with approvals routed to a human. Xena Ugrinsky was an SVP at Booz Allen, global tech lead at KPMG, CTO of RocketDocs, NYU Stern MBA.

What checks out

  • Founder record is real and checkable — the "decades in transformation" claim holds
  • Product genuinely shipped: public pricing, trial, Android app
  • 30+ enterprise integrations (SAP, HubSpot, M365)

What doesn't

  • No technical co-founder or CTO named anywhere
  • "Operating Partner" title — this is a venture-studio placement, not a founder-built company
  • Design partners unnamed; zero customer logos
  • Most contested category in enterprise AI

LiquidRails.ai

liquidrails.ai ↗
6.50/10
UnverifiedNYUdeck said $750KFounders not fully identified

Handles compliant shipping, payments and fulfillment for alcohol DTC e-commerce — a regulated niche with real operational complexity.

What checks out

  • Second-best verified traction in its batch — real named customers
  • Named: Heritage Distilling, Cisco Brewers, Triple Eight Distillery, Grey Sail Brewing
  • Multiple named-customer testimonials
  • Compliance in a regulated category is a genuine moat

What doesn't

  • Founders publish only first names ("Will" and "Ryan") — diligence blocker
  • No press coverage or funding record found
  • Alcohol shipping compliance carries real legal exposure

Nexhelm AI

nexhelm.ai ↗
6.50/10was 7.75 before the refute pass
$20KU RochesterTraction self-reported

Materially narrower than pitched: not a layer over 100+ SaaS tools but an execution layer for wealth management, closing the gap between client meetings and back-office work. Lona Yu (URochester BS, ex-Amazon ML Scientist II on fraud detection) is a Founder-in-Residence at the AI2 Incubator.

What checks out

  • 15 paying firms in 4 months with zero churn (self-reported)
  • SOC 2 Type II achieved
  • Case study: $1.8B AUM RIA, 600,922 records synced, zero failures
  • Founders verified via GeekWire, LinkedIn, Crunchbase

What doesn't

  • Every traction figure is self-reported, none audited
  • Pitch oversells scope: ~20 integrations, not 100+
  • $20K raised is odd for an AI2 Incubator company — ask what it refers to
  • Vertical is filling fast (Zocks, Jump, Powder, Hebbia)

Grego AI

grego.ai ↗
6.00/10was 8.25 before the refute pass
Undisclosed seedNYUdeck said $2MDeck raise unverifiedUniversity tag does not match

Not the general "Security OS" the deck describes — it is an AI auditor for Solidity codebases that synthesizes proof-of-concept exploits. Justus Hanna (alias "Riptide") is an externally checkable top-30 Immunefi bug-bounty hunter; CTO Gregorio Maspero is a maths-olympiad gold medalist. Seed led by cyber-Fund with Vercel CEO Guillermo Rauch angeling.

What checks out

  • Founder ranking is externally verifiable on Immunefi and HackenProof leaderboards, not self-reported
  • Named customers: Centrifuge, Dawn Protocol, Seba, Royco, Tempo, GenLayer
  • Found a flaw putting $27.7M at risk; paid a $250K bounty
  • Ranked #1 AI tool on both Immunefi and HackenProof

What doesn't

  • Round amount never disclosed — the $2M is unverified
  • Pitch scope inflated: "OpenAI for Security" describes a Web3 audit tool
  • No NYU affiliation found anywhere; founders are Miami-based
  • Revenue tied to crypto cycles; crowded field (Trail of Bits, OpenZeppelin, Zellic)

TwinsAI

twinsai.com ↗
6.00/10
UnverifiedNYUdeck said $500KStats self-reported

"Integrated Sales Environment" putting an AI in the room with every seller, scoring calls against methodology in real time. Co-founder Niraj Shah previously co-founded Eventgrid, acquired by DoubleDutch in 2017.

What checks out

  • Repeat founders with a real exit
  • Named logos: Envato, Prefix, Compass, Glasscubes
  • ERA accelerator graduate; demoed at a16z Tech Week

What doesn't

  • The 14% / 11% claims are self-reported vendor metrics with no customer-attributed source
  • Positioning shifted from voice-clone cold calling — possible pivot churn
  • Almost no press footprint
  • NYU tag is affiliation, not founder provenance — these are mid-career operators

Thin

Little survives independent checking. · 16 companies
#CompanyRatingRaised (verified)Deck saysWhat it does
26CRSHMARKET Deep5.75$800K at $10M post$850KLivestream prediction market with 60-second settlement
6Intrinsec AI Deep5.75Unverified$1MAnomaly detection for self-hosted LLMs
28LawVo Comm5.75Not verifiable$1.5MSubscription legal platform with AI copilot
32Scrubin Comm5.75€500K€500KHealthcare workforce network for clinical AI evaluation
11Twinning Labs Comm5.75Unverified$650KDigital-twin simulation of customer and market decisions
14Typework AI Comm5.75Unverified$400KConversational build of SMB business systems
30aperture Comm5.75UndisclosedUndisclosedAI behavioral-interview screening
2ARMAI Technologies Comm5.50$350K$600KAI risk, compliance and audit readiness for banks
17DataRightsai Comm5.50$1.9M$1.9MExpert-trained legal AI for financial-market data rights
9Decode It Comm5.50Unverified$600KColumn-level data lineage and documentation
15Fork Deep5.50$0$0Per-user app customization built by coding agents
35Guidant BioTherapeutics Deep5.50Seeking $20-30M$3MTumor-targeting delivery for solid-tumor cancer therapeutics
22Symitri Comm5.50$5M Series A (May 2024)$6MCookieless advertising via privacy-safe data matching
18MiHu AI Comm5.25Actively raising$0AI contact-center platform across voice and messaging
1Fabrik Comm5.00Unverified$500KOpen data exchange for third-party risk compliance
29Kizuna Comm5.00$1.4M pre-seed$1.4MDecision layer on top of background checks

CRSHMARKET

crshmarket.com ↗
5.75/10was 6.50 before the refute pass
$800K at $10M postRutgersdeck said $850KRegulatory exposureRaise figures inconsistent

"Twitch meets Kalshi" on Monad — trade outcomes of live gaming, esports and IRL events. Co-founder Param Patel is Rutgers-verified; the origin story is a viral college dating show used as the user-acquisition engine.

What checks out

  • Rutgers tie verifies — one of the few that does
  • Live app with active markets and organic crypto-media coverage
  • Genuine distribution instinct: the dating show was the growth hack
  • Strong Daxos thesis fit — digital assets, consumer crypto

What doesn't

  • Regulatory exposure is the story: real-money markets on social outcomes sit outside what Kalshi litigated
  • Resolution integrity — a streamer can influence their own market's outcome
  • Three sources, three different raise numbers ($800K / $750K / $850K)
  • Founders are recent grads with no institutional operating pedigree

Intrinsec AI

intrinsec.ai ↗
5.75/10
UnverifiedStevensdeck said $1MPossible employment conflictSolo founder

Founder Ari Azarafrooz is a Senior Staff ML Scientist at Netskope with NeurIPS/ICML publications and patents. A verified stevens.edu email confirms the school tie.

What checks out

  • The only genuine deep-technical founder in its batch
  • Real publication record and patents in ML security
  • Stevens tie verified via institutional email

What doesn't

  • Appears to be a solo founder — headcount of 1
  • GitHub repos have zero stars; site is soliciting design partners
  • Possible employment conflict — presented as current Netskope staff while founding in an adjacent space
  • Conflation risk with Intrinsec, the established French cybersecurity firm
  • Positioning drift: deck says agent-security amnesia, site sells vLLM anomaly detection

LawVo

lawvo.com ↗
5.75/10was 6.00 before the refute pass
Not verifiableNYUdeck said $1.5MRaise unverifiableSlow trajectory

Toronto-based, founded 2021. Legal AI copilot, document center, 24/7 advice line, lawyer marketplace, plans from $9.99/mo. All three founders verified by name; two are practicing lawyers.

What checks out

  • All founders verified; DigitalOcean published a customer case study
  • Real shipped production system (144 AI agents, 700-800GB legal data)
  • Embedded API is a credible B2B2C channel

What doesn't

  • Five years old and still sub-$2M raised — the trajectory is the finding
  • $9.99/mo consumer legal has no pricing power against LegalZoom and Rocket Lawyer
  • Raise absent from every legal-tech funding tracker that lists peers
  • Infrastructure vanity metrics — no users, revenue or retention disclosed

Scrubin

scrubin.io ↗
5.75/10
€500KNYUPositioning pivotGeography mismatch

Estonian company (Tallinn), live since 2022. Originally a two-way healthcare recruiting marketplace; the clinical-AI-evaluation framing is a recent pivot layered on top.

What checks out

  • Angel money from co-founders of Bolt, Wise, Pipedrive and Testlio — the strongest signal here
  • Real product, live since 2022
  • Funding consistent with public record

What doesn't

  • Positioning drift: HR marketplace (2022) to clinical AI data network (2026) — the pitch sells the newer story
  • 100,000+ "verified" is self-attested; verification method undisclosed
  • Estonia, not UK or NYU
  • No AI-eval revenue evidence

Twinning Labs

twinninglabs.com ↗
5.75/10was 5.50 before the refute pass
UnverifiedNYUdeck said $650KLogo-washingNarrative inflation

NYC, founded 2025, 7-8 employees. Builds digital twin populations from first-party data to test pricing and campaigns. Originally "Twinning AI," a marketing-simulation tool.

What checks out

  • CTO Ankit Das is a verified ex-Microsoft AI engineer
  • NYU tie confirmed via the NYU Entrepreneurship blog

What doesn't

  • Logo-washing: Gojek, Microsoft, Amazon listed as "partnerships" actually track to team members' former employers
  • Narrative inflation — a 7-person pre-seed doing persona simulation now calls itself "a frontier AI lab building enterprise world models"
  • No named customers; "advanced discussions with Fortune 50" is founder-reported
  • Synthetic-persona research carries real validity skepticism; competitor Simile has $100M

Typework AI

typework.ai ↗
5.75/10was 6.25 before the refute pass
UnverifiedNYUdeck said $400KHeadcount inconsistentGeography mismatch

Unifies external comms and internal workflows for SMBs. Founder Henry Yuan holds an NYU Tandon MS in cybersecurity and CS and was previously a Tencent engineer.

What checks out

  • NYU Tandon and Tencent both verify
  • Three named case studies (HK gym franchise, Heep Hong Society, Mana Health)
  • Won BEYOND Expo Innovation Award 2026

What doesn't

  • Headcount claims conflict: 11-50 vs 1-10 for a 2025 company on ~$400K
  • Pitched as USA; every customer, award and partner is Hong Kong, Macao or London
  • Site blocks crawlers — nothing independently checkable
  • Most crowded SMB-AI pitch line of 2026

aperture

aperturehq.ai ↗
5.75/10
UndisclosedStevensNo revenue validationRegulatory exposure

Replaces resume filtering with async video interviews scored on six dimensions, ranked relative to the live applicant pool rather than in isolation. Co-founder Harsh Shroff is a Stevens CS graduate teaching assistant.

What checks out

  • Stevens tie verified
  • Product Hunt launch: 122 upvotes, #7 product of the day
  • Comparative ranking is a real conceptual differentiator

What doesn't

  • Free product means zero willingness-to-pay validation
  • No prior operating pedigree — grad TA, not industry
  • Mercor is running the identical wedge at a multi-billion valuation
  • Regulatory exposure with no visible compliance story: NYC Local Law 144, EEOC, Illinois AI Video Interview Act, EU AI Act

ARMAI Technologies

armai.ai ↗
5.50/10
$350KU Rochesterdeck said $600KDeck raise contradicted

Founded 2024, Charlotte NC. Founders are career banking and compliance executives.

What checks out

  • A real named institutional customer: First Federal Bank of Kansas City, with an on-record quote from their VP of Fintech (March 2026)
  • Fintech Sandbox member; Tampa Bay Wave and RevTech Labs accelerators

What doesn't

  • Outside data shows $350K, not the $600K on the sheet
  • Ordinary pedigree — career banking executives, nothing on the hard-culture list
  • Charlotte NC, not Rochester
  • "Strategic relationship" is vague — may be an LOI, not a paid contract
  • Competes directly with Fabrik in this same session

DataRightsai

datarightsai.com ↗
5.50/10was 7.50 before the refute pass
$1.9MRutgersNo technical leadership named

The deck line ("business intelligence") badly undersells it. Product SAGE interprets contracts, policies and regulations — especially market-data licensing rights — for banks, exchanges and index providers. Vivian Maese was a Latham & Watkins partner who co-chaired their FinTech practice and worked on the SEC Consolidated Audit Trail; co-CEO Ali Hackett spent 30+ years at Citi, CME and LSEG.

What checks out

  • The only raise in 36 companies that independently corroborates
  • FT 2025 Top 20 Innovator; WatersTechnology 2026 Trailblazer
  • Domain-perfect pedigree — founders are the buyer persona
  • Narrow expertise (CAT, market-data licensing) is a genuine moat

What doesn't

  • No CTO or technical leader named anywhere — material gap for an AI product
  • No disclosed customers or ARR
  • Rutgers tie appears in neither founder bio
  • Both founders late-career; Maese also runs maesellc.com — confirm full-time

Decode It

decodeitnow.com ↗
5.50/10was 5.75 before the refute pass
UnverifiedU Rochesterdeck said $600KCompetes with a free native feature

Automates data documentation and builds column-level lineage graphs. In beta, Snowflake only. Founder Irina Ivanova holds a URochester Simon MBA and came from Gartner.

What checks out

  • URochester tie verifies (Simon MBA)
  • Appears on f4.fund's portfolio page — an institutional check likely exists

What doesn't

  • Snowflake ships native column-level lineage in Snowsight — a free competitor inside the only warehouse they support
  • Technical co-founder invisible; founder pedigree is business-analyst
  • No traction published at all
  • Category is brutally crowded and better capitalized (Atlan, Alation, Collibra)

Fork

withfork.co ↗
5.50/10
$0RutgersUniversity claim failsNon-technical founder

An SDK letting a SaaS vendor mark surfaces "forkable" so each user's agent-built features run isolated to their own instance. First shipped artifact is a free Chrome extension over Gmail and Calendar.

What checks out

  • Genuinely interesting wedge worth tracking
  • Real shipped artifact; YC Fall 2026 applicant

What doesn't

  • Rutgers claim fails — founder Paul Demick's verified education is Florida State and Wake Forest, in accounting
  • Non-technical founder on a deeply technical product (agent-authored code merged into production SaaS)
  • Near-zero traction: Show HN got 5 points, 2 comments
  • YC already funded a direct competitor (Vendo)
  • Adoption requires vendors to grant users code-execution rights inside their own app

Guidant BioTherapeutics

guidantbio.com ↗
5.50/10
Seeking $20-30MRutgersdeck said $3MNot a seed companyIP assigned elsewhere

Not a seed startup. This is the asset spinout of Sonnet BioTherapeutics (formerly NASDAQ: SONN), created after Sonnet's public shell was reverse-merged with a crypto company. The underlying science is John K. Cini's FHAB albumin-binding platform.

What checks out

  • Real clinical history: 103 patients dosed across three legacy Phase 1 studies
  • Substantial patent estate including granted US12006361B2
  • Platform is genuinely clinically de-risked

What doesn't

  • Pitched as a $3M Rutgers seed company; is actually a public-company carve-out seeking $20-30M
  • Every patent is assigned to Sonnet BioTherapeutics, not Guidant — chain of title after a reverse merger is the key risk
  • The bladder-cancer asset actually being pitched (GDT-002) is IND-ready, not clinical
  • Openly seeking a partner to advance it — reads as inability to fund alone
  • IL-12 has a long history of clinical toxicity failures

Symitri

symitri.com ↗
5.50/10
$5M Series A (May 2024)NYUdeck said $6MTwo deck numbers inflatedStale round

Founded 2023 in NYC by ad-tech veterans with a formal board including Jeff White and Peter Bordes. Not a university startup.

What checks out

  • The 5.6x conversion lift is tied to a named client (TRUBAR) in a published case study
  • Institutional board and IAB Tech Lab / ANA membership
  • Experienced ad-tech operating team

What doesn't

  • Deck says $6M; every public source including Symitri's own press release says $5M
  • Deck says 53% lower CPA; their own website says 48% reduction in CAC
  • Series A closed May 2024 — over two years with no follow-on in a hot category
  • Named co-founder David Kohl is absent from the current team page

MiHu AI

mihu.ai ↗
5.25/10
Actively raisingStevensdeck said $0Traction claims uncorroborated

Legal entity "Hunters AI, Inc" at a Delaware virtual address; operations appear to be in Istanbul. The automotive framing is a pitch-day narrowing — it sells into seven verticals.

What checks out

  • Stevens MS in CS verified for co-founder Volkan Demir
  • Product is genuinely live across multiple channels
  • Released an open-source Turkish speech-to-text model

What doesn't

  • Claims 500+ businesses and 10M+ conversations with zero independent corroboration anywhere — a company at that scale leaves a trail
  • No named team member on the website despite the scale claims
  • Delaware virtual address paired with Istanbul operations while pitching as USA
  • Entity-name sprawl: MiHu / Hunters AI / MindHunters AI; collides with $118M-funded Hunters.security
  • Commoditized category: Numa, Impel, Fullpath, Cerence

Fabrik

www.thetrustfabrik.com ↗
5.00/10
UnverifiedRutgersdeck said $500KCrowded category

"Plaid for Enterprise Third Party Risk" — cuts a 4-6 week vendor compliance chase to 12 minutes. Co-founder Steve Francolla is Rutgers-verified, ex-Amazon Ads, and holds two patents.

What checks out

  • Rutgers tie verifies; founder holds two patents
  • Third Party Risk Association incubator member with a demo-day slot

What doesn't

  • Pedigree is corporate product/BD, not hard-culture engineering
  • No named customers, no revenue evidence
  • Worst same-name conflation in the cohort — at least five unrelated "Fabrik" entities
  • Extremely crowded: Vanta, Drata, SecurityScorecard, Whistic, SafeBase
  • Competes directly with ARMAI in this same session

Kizuna

kizuna.solutions ↗
5.00/10was 7.50 before the refute pass
$1.4M pre-seedNYUdeck said $1.4MPre-traction

API-first decision layer, not a screening vendor — scores adjudication decisions against job-specific risk factors and criminology research. Megan Goddard (NYU 2016) was at Checkr 2018-2022 in product compliance; CTO Igor Vokatch-Boldyrev was a founding engineer at Checkr; CLO Tyler Browne came from Checkr and Apple.

What checks out

  • Raise verifies exactly: $1.4M pre-seed, April 2025
  • Founding team came out of the incumbent they are now competing with
  • Deliberate liability design — employer stays the decision-maker
  • Live product with FCRA compliance tooling and 200+ policy templates

What doesn't

  • No named customers, no revenue on public record
  • Checkr can ship this as a feature
  • TAM constrained — a decision layer on someone else's data
  • Weak Daxos thesis fit: HR compliance, not digital assets or frontier tech

Pass

Claims contradicted, or nothing verifiable exists. · 11 companies
#CompanyRatingRaised (verified)Deck saysWhat it does
24Galya Comm4.75$0$0"Taste graph" context layer for AI personalization
10ensue Deep4.50Unverified$1MAgent swarms that design and run ML experiments
25Claylight Comm4.25Unverifiable$300KAgentic website and portal builder
3Diligent4 Comm4.25$0$0Governed AI execution context for high-stakes orgs
8TraceIQ Comm4.25$0$0Cost observability for AI agent runs
31Velric Comm4.25$0$0AI-evaluated work missions for talent assessment
5Keyboard Deep4.00Unverified$1.5MGovernance layer for how teams connect data to AI tools
27Mori Deep4.00No record found$2MCovered calls sold to market makers on Solana
21BizMeet Creator Comm3.75$0$0Operating system for brand-creator programs
4CognitivTrust Comm3.50$0$0System of evidence for AI agent actions
34Trinity Aesthetics Comm2.50Unverifiable$500KAI revenue discovery inside medical practices

Galya

www.galya.io ↗
4.75/10
$0NYUCredit programs presented as backing

Models human taste and affinity as structured context that AI agents can query. Explicitly in beta.

What checks out

  • Berkeley SkyDeck affiliation
  • Building in public with a visible community

What doesn't

  • Lists AWS Activate and NVIDIA Inception alongside investors — those are free credit programs, which is misleading
  • Zero named customers; pre-revenue, pre-product
  • No verified founder pedigree
  • Berkeley SkyDeck conflicts with the deck's NYU tag
  • Highly abstract thesis

ensue

ensue.dev ↗
4.50/10was 7.00 before the refute pass
UnverifiedNYUdeck said $1MPositioning churnPre-traction

Legal entity Mutable State Inc. You supply a dataset and an evaluation; agents propose strategies, run experiments and return the best model. CEO Austin Baggio came from Google, NEAR Protocol and o1Labs/Mina; CTO Sai Vegasena came from Trail of Bits and Zoom.

What checks out

  • Best-verified engineering team in its batch — Trail of Bits is elite security engineering
  • Founders' NEAR/Mina/o1Labs history is a real digital-assets adjacency
  • Substantial self-reported benchmarks: 10,000+ experiments, 3,100 coordinated training runs

What doesn't

  • No customers, named or otherwise
  • Positioning churned completely — was "shared memory layer for agents" a year ago
  • Raise unverified; do not credit the "$1.5M Mutable seed" (different company)
  • AutoML plus agentic research is capital-heavy and competes with frontier labs

Claylight

claylight.ai ↗
4.25/10was 5.25 before the refute pass
UnverifiableNYUdeck said $300KNo founder identifiablePitch oversells product

Not what the pitch says. The production bundle shows an agentic website builder — give it a brief, the agent researches, picks a theme and populates modules. Paid tiers gated by "space limits." That is a Linktree/Framer-class product, not "a Personal Language Model that runs your entire online presence."

What checks out

  • Real shipped software: React SPA on Convex, live paid-tier enforcement, admin surface
  • Someone competent built this

What doesn't

  • Not a single founder is named anywhere — no team page, no LinkedIn, no Crunchbase, no press
  • Pitch language materially oversells what the code does
  • No users, revenue, press or reviews
  • Saturated category: Framer AI, Durable, Wix ADI, Lovable, Linktree
  • Entity identification is probable, not confirmed

Diligent4

diligent4.com ↗
4.25/10
$0NYUName collides with $1B incumbent

Tel Aviv, 2-10 people. Founder Tslil Aloni holds an NYU economics PhD; his prior venture was a Madrid property-management company.

What checks out

  • NYU PhD verifies — the university tie is real
  • Founder is a genuine quantitative economist

What doesn't

  • Severe conflation with Diligent Corporation, the $1B+ GRC incumbent selling to the same buyer
  • Product-name inconsistency: Diligent4 vs Flingoos, whose domain does not resolve
  • Zero traction, no team page
  • Founder is an economist with no engineering or deep-tech operating history

TraceIQ

traceiq.dev ↗
4.25/10was 5.00 before the refute pass
$0NYUFounders unverifiable

Per-run spend broken down by token, tool call, retry and workflow. Coherent wedge, but the site is not indexed by search engines and was found only by direct domain probing.

What checks out

  • Product thesis is coherent and matches the pitch
  • Simple onboarding (add a SKILL.md file, send traces)

What doesn't

  • Not a single founder credential is verifiable — one name from a booking link, no LinkedIn, no pedigree
  • Zero third-party footprint: no Crunchbase, Product Hunt, GitHub org or press
  • Site lists an email at a domain they do not own — template boilerplate left in the legal pages
  • Six-plus same-name entities, one in the identical AI-agent category
  • Crowded by funded incumbents: Langfuse, Braintrust, Helicone, Arize, Datadog

Velric

velric.ai ↗
4.25/10
$0RutgersClaims internally inconsistentLogo-wall risk

CEO Mahir Laul is billed in a podcast as an 18-year-old NYU dropout trained in acting at Interlochen Arts Academy. CTO Sara Pedron has no published background.

What checks out

  • Published pricing is real and specific ($5,000 pilot, $15,000/yr)
  • Genuine hustle — visible media placement effort

What doesn't

  • Homepage claims 1,000+ recruiters; founder interview says dozens of companies onboarding — internally inconsistent
  • Logo wall names Eton College, CUNY and SeedLegals with none independently confirmed — highest-risk claim to diligence
  • Rutgers tag contradicts every public source (NYU)
  • No technical or domain pedigree behind an AI-evaluation benchmark claim

Keyboard

keyboard.dev ↗
4.00/10was 6.75 before the refute pass
UnverifiedNYUdeck said $1.5MSecond founder unidentified

Gives IT and Security visibility and control over how teams connect internal apps and data to AI tools (Claude, Cursor, ChatGPT) and the agents they build. Co-founder Stephen Roland is ex-Stripe with an NYU Stern MBA.

What checks out

  • Only company in its batch hitting a named Daxos criteria employer (Stripe)
  • Ships real code: keyboard-local MCP server at 174 GitHub stars
  • NYU Stern MBA, Co-Chair NYU Entrepreneurs Network — university tie is real

What doesn't

  • Stripe role appears to be partnerships, not engineering — confirm on a call
  • Co-founder #2 is not identified anywhere
  • "Keyboard" is effectively unsearchable — real distribution handicap
  • Competitor Geordie (Snyk/Veracode/Darktrace founders) won RSAC 2026 Innovation Sandbox

Mori

mori.finance ↗
4.00/10was 6.25 before the refute pass
No record foundRutgersdeck said $2MRaise likely misattributedCustody contradicts pitch

A genuinely shipped Solana covered-call app — verified by pulling the production bundle. Phantom wallet, European-style options, fully collateralized, settlement at 08:00 UTC, RFQ to liquidity providers. Founders "Pavan" and "Jason" previously founded Yomi Games.

What checks out

  • Product is real and non-trivial — verified in the shipped JS bundle
  • The "sell to market makers" claim checks out: genuine RFQ to liquidity providers
  • Best Daxos thesis fit in its batch — Solana DeFi structured products
  • Answerable in one call — the open questions are specific

What doesn't

  • The $2M appears to be Yomi Games' 2021 round, not Mori's — no record of a Mori raise exists
  • "Natively on Solana" contradicts their own docs: assets sit with an unnamed institutional custodian in alpha
  • Not on DefiLlama — no verifiable TVL or volume; no audit published
  • Founders publish no surnames while holding user funds custodially
  • Category graveyard: Friktion and Katana ran this exact product and wound down

BizMeet Creator

no website found
3.75/10
$0RutgersNo website

No standalone website could be found — only a LinkedIn company page.

What checks out

  • Founder Kristina Knezevic is named and holds a CPA credential

What doesn't

  • No website exists
  • No traction, no product evidence, no named customers
  • Superlative-heavy copy ("ultimate command center", "exponential ROI") with nothing behind it
  • Multiple unrelated "BizMeet" entities — confirm which is pitching
  • No prior employers verified

CognitivTrust

www.cognitivtrust.com ↗
3.50/10
$0NYUSOC 2 claim implausibleMultiple inconsistencies

Waitlist-stage. CEO Yogesh Trivedi has an NYU credential and a 15-year GRC career including the Federal Reserve Bank of Chicago.

What checks out

  • NYU tie verifies; CISA and CISSP credentials are real

What doesn't

  • The domain previously hosted an unrelated job-board blog — repurposed, not built
  • Three inconsistent positionings across LinkedIn, the website and this deck
  • Claims SOC 2 Type II with SSO, RBAC and data residency while at $0 raised and pre-launch — this does not hold up
  • Claims 11-50 employees at $0 raised — internally inconsistent
  • Only one person is identifiable; no team page anywhere
  • Menlo Park, not NYC

Trinity Aesthetics

no website found
2.50/10
UnverifiableU Rochesterdeck said $500KZero public footprint

No domain resolves. No company website, LinkedIn page, entity record, customer, press or patent could be located across four search engines and Google Patents.

What checks out

  • nothing verified

What doesn't

  • Complete absence of public record despite claiming to be in-market with $500K raised
  • No founder names found in any source
  • Nothing verifiable exists to underwrite
  • Name collides heavily with medspa clinics and Trinity Health