Ollama Just Raised $65M With 14 People. The llama.cpp Contributor Graph Is the Map.
Ollama's $65M Series B funds a hiring sprint from 14 people. Here's the Docker alumni, llama.cpp, and vLLM sourcing map recruiters should build now.
On July 9, 2026, Ollama closed a $65M Series B led by Theory Ventures, bringing total funding to $88M. Founder Jeff Morgan told TechCrunch the platform now serves 8.9 million monthly developers across 85% of the Fortune 500 with only 14 employees. If you are a recruiter or founder trying to hire against them, or hire from them, the "Docker alumni" LinkedIn search everyone will run this week is the wrong map.
The 14-person leverage number is the actual pitch
Ollama has roughly 636,000 monthly developers per employee. That ratio beats every open-model runtime competitor on capital efficiency, and it is what recruiters should be quoting to candidates instead of the funding total. Inferact (the vLLM commercialization company) raised $150M in seed in January 2026. Together AI just closed $800M. Replicate is bigger. Ollama is smaller than all of them and reaches more developers than any of them.
That leverage also tells you exactly why the hiring bar will be brutal. Fourteen people cannot afford a bad senior hire. Morgan and Michael Chiang will index hard on people they already know, people their angels vouch for, and people whose GitHub history proves they can ship at the layer Ollama sits on. Which is where the sourcing work actually starts.
The Docker alumni pool is already skimmed
The obvious move is to pull the "current or past: Docker" alumni list on LinkedIn and start there. Do not bother. Solomon Hykes, Docker's founder, is on the Series B cap table as an angel. So is Marianna Tessel, former Docker EVP, now on the Cisco board. The senior Docker network is already aligned with Ollama, which means the highest-signal veterans are either already contacted, already investing, or already committed elsewhere in Hykes' portfolio.
The productive move is one hop out. Three specific lanes:
Kitematic-era colleagues, not Docker Inc. hires
Morgan and Chiang built Kitematic before Docker acquired it in 2015 and folded it into Docker Desktop. Kitematic was a tiny team. The people who worked directly with them then, and the early Docker Desktop PMs and engineers who inherited that codebase between 2015 and 2019, are the actual "worked with the founders" pool. That is maybe 30 to 60 people globally, not the 3,000+ that a "past company: Docker" filter returns.
Post-2019 Docker Enterprise / Mirantis splits
When Mirantis bought Docker Enterprise in November 2019, Docker fractured into two orgs. The engineers who stayed on Docker Desktop after that split have the most relevant modern context: cross-platform desktop app for developers, freemium-to-paid conversion, and the exact OSS-to-commercial tension Ollama is walking into. Everyone else went to Mirantis, Rancher, or cloud-native infra jobs. Both sub-populations are worth mining, but the profiles you want lean toward the Desktop-side survivors.
Waterloo CS
Both founders came out of the University of Waterloo. Their strongest referral pipeline will run through Waterloo CS alumni scattered across Stripe, Cloudflare, Shopify, YC W21 through W23 cohorts, and Docker itself. A LinkedIn title search for "Docker" will not surface this. You need to search by school, graduation window, and current company, then filter by GitHub activity in Go or Rust systems code. This is where general-purpose sourcing tools fall over and why we built Refolk: you describe the person in plain English ("Waterloo CS grads at YC-backed infra startups who contribute to Go container tooling") and get a ranked shortlist across GitHub, LinkedIn, and the open web.
The real map is the llama.cpp contributor graph
Here is the insight that most recruiters chasing this round will miss. Ollama is not really a Docker company. It is a llama.cpp company with a Docker-style UX layer on top.
llama.cpp was started in March 2023 by Georgi Gerganov as a pure C/C++ implementation of Llama inference with no dependencies. It is the de facto core of almost every serious local inference tool shipping today, including Ollama and LM Studio. The people who send meaningful pull requests to ggerganov/llama.cpp already understand GGUF quantization, Metal and CUDA kernels, KV cache management, and the exact codebase Ollama depends on. They are also, almost by definition, the people Inferact and LM Studio (Element Labs) want to hire.
Ollama is not a Docker company with an AI layer. It is a llama.cpp company with a Docker-style UX.