AI Venture Engineer

Please do not send a CV. Send me something you built instead. A repo. Something you deployed. An experiment that failed in an interesting way. Here is why. Any LLM writes a convincing CV in nine seconds, and I will get a hundred of them. A CV proves you can get a model to say what I want to hear. It does not tell me whether you can build, or whether you can pick things up fast enough to build soon. Working code tells me both. ABOUT TECHTRUTH Most pitch decks mention AI. TechTruth helps investors see how much of that is real for a specific company. Per deck the product runs hundreds of background checks, benchmarks the result against the 1,300 decks already in the system, and folds in the feedback we get from founders and investors who disagreed with our score. The output is a score an investor can actually argue with. Where we are now: - 1,300 pitch decks fully analysed - 20 VC and PE funds using it - 400 founders who have pushed back on our conclusions The product is live and in use — and now ready to be made much better. That is what you would work on. TechTruth was started by Why Commit Capital, an early-stage investor in AI, climate and deep tech. Enjins is our build partner: an AI engineering agency that has run 100+ technical due diligences for investors since 2018. Volve Capital is our implementation partner: an operating VC fund where TechTruth runs inside a live dealflow. WHAT IS NEXT FOR THE PRODUCT - Better scoring — the model is directionally right and not yet precise. - Sharper benchmarks — a deep-tech seed deck should not be judged like a B2B SaaS Series A. - More sectors — each one needs its own reference points before the score means anything. - A self-service portal — so a fund can run a deck without us in the loop. - Rollout to more funds and incubators, which changes what the product has to handle. - Higher quality throughout — fewer false flags, clearer reasoning, faster turnaround. You are junior and you will not be alone. Experienced engineers at Enjins and investors at Volve and Why Commit are around you every week. But nobody is going to hand you a spec. WHAT YOU WILL DO Build. Improve the product: evaluation flows, scoring models, and the Ghost Code Scanner we ship with Enjins, which checks a startup's GitHub activity without anyone sharing private code. You write Python that other people have to read, you work in Git, and you get things into production rather than into a notebook. Modern LLM tooling, room to experiment — but what ships has to keep running. Measure. Run real decks through the system and find out where it is wrong. Build the sector benchmarks. Test whether the model genuinely reproduces expert judgement. That last part is unproven, and proving it is a real part of the job. Work the dealflow. Look at incoming companies with us: does the technology hold up, and does the claim match what has actually been built? You form a view and you say it out loud, with senior people in the room to argue back. Demo days and investor sessions are part of it, and what you hear there goes straight back into the product. WHO WE ARE LOOKING FOR An AI-native builder. Someone curious enough to take things apart, not someone waiting for a ticket. What matters most: you build with AI because you cannot really help it. Half-finished agents, a scraper you wrote to settle an argument, a fine-tune you ran at two in the morning to see what would happen. That tells us more than any transcript. What you can already do: Python you are comfortable in, Git as a reflex, and enough understanding of how software gets tested and deployed that CI/CD and basic MLOps are not new words to you. LLM APIs, prompting, evaluation. You have shipped something — a course project, a side repo or a hackathon demo all count. Running code beats a perfect plan. What you will pick up here: getting machine learning into production properly. Pipelines, monitoring, evaluation you can trust, cloud. You will learn this next to Enjins engineers who do it for a living. Nobody expects you to arrive with it. About diplomas: a degree in AI/ML, data science or computer science is one route here. It is not the only one and it is not required. We care about what you have made. About uncertainty: where this product goes in two years is not decided, and neither is where this market goes. If you need a fixed roadmap to feel safe, you will not enjoy this. If that is the interesting part, we should talk. How you work: you end up in rooms with founders and investors, and you ask when something does not add up. Curiosity is the job. WHAT YOU GET - Real users and real data from day one. Not a sandbox and not a made-up brief. - Production engineering learned next to Enjins engineers, and a live fund at Volve to see your work land in. - Regular contact with funds, founders and incubators across the Benelux and DACH. You meet companies before the market does. - Room to co-invest alongside Why Commit on deals you helped look at. Same terms as us, no minimum ticket theatre. - Hours that fit around you, part-time or full-time, with a base in Amsterdam. WHERE YOU WILL WORK Your base is Amsterdam. Expect roughly one or two days a week somewhere other than your base, and expect that mix to shift as the work does. You will work together with tech team of www.enjins.com and/or implementation partner www.volve.capital. - The Stack, Amsterdam — your home base. Opening September 2026 as a workspace for AI builders: 4,500 m2, 200+ founders planned by 2029. Co-founded by Maarten Stolk (Deeploy, Enjins). - Enjins HQ, Utrecht — where most of the engineering happens, and where the experienced engineers sit. - Merantix AI Campus, Berlin — Enjins' German base and one of Europe's densest AI clusters. Occasional, not weekly. HOW TO APPLY Do not send a CV. Send your work. A repo, something you deployed, or a write-up of an experiment that failed in an interesting way. Tell us what you built with AI, what it taught you, and how many hours a week you have available. Email: bastiaan@whycommit.com We read everything and reply within a week. The first conversation is a call. The second is a working session on something real — no take-home assignment invented for the occasion.