Venture Capital · AI Research · San Francisco

Yigit
Ihlamur

The VC who publishes research papers, scans 300+ startups a week by AI, and still closes the check in days - not months.

Vela Partners Oxford · Google AI Research NeurIPS 2025
Yigit Ihlamur - Co-founder and General Partner at Vela Partners
40+ AI startups
backed
50+ Research
papers published
300+ Startups
scanned weekly
2-7 Days to soft
commitment
58 Academic
citations
#1 Product Hunt
of the Day

The man who turned gut feel into an algorithm

His father carried a computer on his shoulder up five flights of stairs every night. No elevator. A technology entrepreneur in Turkey in the early 1990s, balancing a day job and a startup dream. Little Yigit waited at the top. The computer arrived. And something stuck.

That image - a man willing to haul the future home one floor at a time - lives in everything Yigit Ihlamur has built since. Today, as co-founder and General Partner of Vela Partners, he's the one carrying the weight. Except now the cargo is a proprietary AI system that reads three hundred startups a week, distills them into signal, and helps him write a check faster than most VCs schedule their first call.

Vela Partners is not a conventional venture firm. Yigit describes it as an AI-native quantitative VC - a phrase that would sound like marketing if the research didn't back it up. Backed by tier-one VCs and armed with what Yigit calls Ventech (venture technology), the firm deploys $100K-$500K seed checks with soft commitments in two to seven days. Their AI suite filters millions of companies down to the 300-odd worth a second look each week, with a human - Yigit and his team - acting as the final filter.

"As early investors, the founder matters more than anything else to us."

- Yigit Ihlamur, Co-founder, Vela Partners

He started at four. DOS games, shell commands, an introduction to computing that most kids today get from YouTube. By his teens, his father had taught him Turbo Pascal and pulled him into client meetings, giving him a front-row seat to the mechanics of running an enterprise technology company. His mother, a mathematician, held the family together financially while his father built - a quiet lesson in what conviction looks like from the inside.

Chess came next. Not casually - he ranked in Turkey's national top ten. The game teaches something that holds in venture capital: patterns repeat. The position on the board at move 12 tells you something about move 40. Most people don't read that far ahead. Yigit learned early to try.

He took that pattern-recognition discipline to Koç University, where he studied Industrial Engineering, and then to Oxford, where he completed an MSc in Computer Science with a focus on AI and machine learning, writing a master's thesis on recommendation engines while living at Kellogg College. Oxford, he says, gave him exposure to "pioneering individuals and ideas" - an understatement for a program that shaped how he thinks about machine judgment at scale.

Five years at Google followed. Senior Program Manager, first in Dublin working on EMEA business products, then in Mountain View on Google Cloud and Enterprise. He applied his Oxford ML training daily - assessing technology, shipping product features, navigating the kind of organizational complexity that only a company at Google's scale can produce. He calls that period foundational: "Having operator experience builds founder rapport." He learned how products break before he started funding the people building them.

"Speed wins over perfection - it's that simple."

- Yigit Ihlamur

The pivot came after Google. Six months of exploring ideas. A group of entrepreneurs he kept running into who, he noticed, preferred sharing networks and knowledge over talking about their jobs. That cluster became the seed of Vela Partners, co-founded in 2017 with Fuat Alican. The premise was almost counterintuitive: use AI to invest in AI. Not as a gimmick - as a genuine edge.

The numbers bear it out. Vela has backed 40+ AI startups including Limitless (AI wearable for recording personal information), Grabango (checkout-free retail), Interviewing.io (bias-reduction hiring), Base Operations (urban safety navigation), Meditopia (mindfulness), Bito (developer tools), and Vieu. The firm is a board observer or member at five portfolio companies. Their Vela Terminal product hit #1 Product of the Day and Month on Product Hunt.

But what separates Vela from other AI-focused VCs is the research program. Yigit is not just an investor who reads papers - he writes them. His Google Scholar profile lists 58 citations, an h-index of 5, and a body of work including Founder-GPT (applying self-play techniques to evaluate founder-idea fit), Automating Venture Capital (LLM-powered segmentation and labeling of founders), and the Startup Success Forecasting Framework, which won a Best Poster award at NeurIPS 2025. The workshop was about AI; the poster was about predicting which startups survive. The irony is deliberate.

He reads Sapiens and F.S.C. Northrop's The Logic of the Sciences and the Humanities. He has practiced Benjamin Franklin's early-rising discipline for seventeen consecutive years. He lives in the Bay Area with his spouse and two children. He values, above all things, being a good human being and working unbelievably hard. The rest, he says, follows.

Vela Partners is a Pledge 1% member, supporting women-focused entrepreneurship and scholarships. The firm's investment criteria aren't just a checklist - they're a philosophy. Why does this founder exist for this problem? Where has mastery shown up before? How fast do they move? What's the "why now"? These questions, filtered through AI and human judgment together, are what Ventech looks like in practice.


Five filters. Zero exceptions.

Yigit doesn't use a gut feel checklist. He uses a framework built from pattern recognition across hundreds of founders - distilled into five criteria that must all pass before Vela writes a check.

1
🧭
Why
Personal motivation rooted in lived experience - not just market opportunity
2
🎯
Mastery
Demonstrated excellence in at least one domain before this startup
3
Speed
Fast decisions, fast execution. Speed wins over perfection, always.
4
💡
Idea
A clear, exciting articulation that attracts world-class talent
5
Timing
A compelling "why now" that competitors can't simply copy today

What Ventech looks like

Venture technology: proprietary AI replaces the gut feel of traditional VC with quantitative signals, while humans remain the final decision layer.

01
AI Scan
Millions of startups filtered weekly to 300+ signals
02
Founder Score
LLM-powered segmentation & feature engineering
03
Human Filter
Team acts as human-in-the-loop to decide
04
Fast Decision
Initial decision in minutes. Soft commit in 2-7 days
05
Network + Check
$100K-$500K + 5-100 tier-1 introductions

What Yigit actually says

We're not in the business of educating entrepreneurs, we're in the business of choosing promising teams.

Entrepreneurs need to have a personal story and motivation behind why they do what they do.

Founders who achieved mastery in their prior careers and personal lives are curious and self-driven, and have grit.

I value being a good human being and working unbelievably hard. The rest follows.

Startup-life is hard. Really hard. The entrepreneur needs to have a history of excellence in a chosen field.

Once I had that long-term oriented view, I stopped feeling overwhelmed.

From DOS games to quantitative VC

~1990s - Early
Introduced to computers at age four via DOS games in Turkey. Father teaches him Turbo Pascal. Begins competing nationally in chess, reaching top-10 in Turkey.
~2005-2008
Studies Industrial Engineering at Koç University, Istanbul - one of Turkey's premier research universities.
2009-2010
MSc Computer Science (AI/ML focus) at University of Oxford, Kellogg College. Writes master's thesis on recommendation engines. "I ultimately chose Oxford due to its centuries-long reputation in innovation."
2010-2015
Senior Program Manager at Google - first in Dublin on EMEA business products, then Mountain View on Google Cloud and Enterprise. Applies Oxford ML training to real-world product features at scale.
2015-2017
Leaves Google. Spends six months exploring startup ideas. Connects with entrepreneur community in Silicon Valley whose members prefer sharing networks over day jobs. Seeds the idea for Vela.
2017
Co-founds Vela Partners with Fuat Alican. Pioneers "Ventech" - the application of AI and data science to venture capital decision-making. Begins building proprietary startup evaluation algorithms.
2023-2024
Publishes a wave of AI research: Founder-GPT, Automating Venture Capital, GPTree, and Startup Success Forecasting Framework. Collaborates with Oxford researchers.
2025
Startup Success Forecasting Framework wins Best Poster at NeurIPS 2025 Workshop. Vela Terminal launches to #1 Product of Day/Month on Product Hunt. Leads $6M Bito seed round.

Where the pattern began

University of Oxford
MSc Computer Science - Artificial Intelligence & Machine Learning
2009 - 2010 · Kellogg College

Thesis: Recommendation Engines. Gained exposure to machine learning research that directly shaped Vela's quantitative approach to startup evaluation.

Koç University
BSc Industrial Engineering
Istanbul, Turkey

Systems thinking, optimization, and operational research - foundations that translate directly to building investment processes at scale.

Companies that passed the five filters

Limitless
AI wearable that records, transcribes, and searches your personal history
Grabango
Checkout-free shopping technology for brick-and-mortar retailers
Interviewing.io
Bias-reduction platform for technical hiring at scale
Meditopia
Mindfulness and mental wellness application · Board member
Bito
AI-powered developer tools · Seed round $6M (May 2025)
Vieu
Seed round $11M (Oct 2024)
Axiom Cloud
Series A $7M (Jan 2023) · Energy optimization for commercial refrigeration
Base Operations
Urban safety and security navigation for enterprise teams

The VC who publishes the playbook

Yigit doesn't just apply AI to investing - he contributes to the field. His research explores how language models can evaluate founders, predict startup outcomes, and automate the most human parts of venture capital.

Total citations: 58 · h-index: 5 · Verified at Google Scholar

Things that don't fit the deck

He was introduced to computing at age four through DOS games - before he could read properly. His father taught him to type commands before he could type sentences.
His mother was a mathematician who held the family together financially while his father built his startup. Yigit credits her conviction as equally formative to his father's risk-taking.
Top-10 national chess player in Turkey. The game that trained him to read pattern sequences 30 moves ahead now helps him read founder trajectories 30 months ahead.
He has practiced Benjamin Franklin's early-rising discipline for 17 consecutive years without a gap. Discipline, not motivation.
His Oxford master's thesis was on recommendation engines - a technology that now underpins everything from Netflix to the LLMs he uses to score founders.
Vela's AI scans and surfaces 300+ early-stage startups from among millions every single week. The team then acts as the human-in-the-loop. Speed with judgment.
His handle across LinkedIn, Instagram, HuggingFace, and GitHub is consistently "yihlamur" - one character shorter than the name, which somehow perfectly describes his operating style.
Vela Partners is a Pledge 1% member, committing a portion of equity, time, and product to support women-focused entrepreneurship and scholarships globally.
Share this profile LinkedIn Twitter / X Facebook