See who is building, and what, from the commits that land months before the deck.
Founders who build alike
ML serving researcher heading RadixArk
Leads LLM serving and evaluation at RadixArk, after Member of Technical Staff at SpaceXAI. Contributes to lm-sys/FastChat, a platform for serving language models, and FMInference/FlexLLMGen, a tool for running language models on a single GPU.
CRACKED
Active 16 days ago
Stanford, United States
Leads LLM serving and evaluation at RadixArk, after Member of Technical Staff at SpaceXAI. Contributes to lm-sys/FastChat, a platform for serving language models, and FMInference/FlexLLMGen, a tool for running language models on a single GPU.
251 contributions this year
Core LLM inference maintainer building serving tooling at Inferact
Builds LLM inference and serving systems at Inferact. Maintains vllm-project/vllm, an engine for large language models, and contributes to pytorch/pytorch, a library for neural networks in Python.
CRACKED
Active 1 month ago
Beijing, China
Builds LLM inference and serving systems at Inferact. Maintains vllm-project/vllm, an engine for large language models, and contributes to pytorch/pytorch, a library for neural networks in Python.
2,000 contributions this year
CEO cutting LLM inference costs through open source at Inferact
Leads inference cost reduction at Inferact, with a Member of Technical Staff background from Character.AI. Works on vllm-project/vllm, a serving engine for LLMs, and ucbrise/clipper, a prediction-serving system.
CRACKED
Active 14 days ago
Berkeley, United States
Leads inference cost reduction at Inferact, with a Member of Technical Staff background from Character.AI. Works on vllm-project/vllm, a serving engine for LLMs, and ucbrise/clipper, a prediction-serving system.
2,966 contributions this year
For VCs, scouts and platform teams.
Step 1
Describe who you want in Claude, or send it to the API: a place, a company, a repo or the kind of work.
Step 2
Each person comes back with the repos they own and contribute to, their contribution garden, a cracked score and a one-line why.
Step 3
Keep a shortlist, write the fields to your CRM, or build them into your product over the API.
Run it in Claude, or build on the API. See how Vamo compares with GitHub search.
It reads what people ship on GitHub alongside their LinkedIn history. A Stealth title with fresh commits shows who is building, and the repos show what.
Yes. Filter by where people worked, when they left and where they studied, then rank them by what they build.
Yes. Deep research reads a person's code, collaborators and history and writes it up as one report.
No. Ask in plain words from Claude, ChatGPT, Cursor or any AI agent. Engineers who want to build on Vamo can call the API directly.
Contribution calendars are read from GitHub and refreshed every day, and Discover lists are rebuilt every hour. A founder's new commits show on their Vamo profile within a day.
Book a call. We'll run a search on a space you're watching.
© 2026 Vamo Holdings, Inc.
