The posting
Your Role & How we work
We are looking for an AI Engineer to build the AI systems at the heart of BIT Capital's investment process. Aion — our agentic research platform with first-class access to our data, models, and tools — is in production and used by the investment team every day. Around it, a growing family of AI systems generates investment ideas, performs research, monitors tail risks, and distills millions of documents into insight. You will build both the platform all of this runs on and the AI research products themselves. Our stack is Python, SQL, AWS, and Databricks, with models and agentic frameworks from the leading AI labs.
This is a deep engineering role at the frontier of applied AI. The craft we care about most is the harness: turning fuzzy, judgment-heavy requirements into machinery around agents — context, tools, specs, verification, feedback loops — that produces outcomes that decision makers trust and get better with every iteration. As an individual contributor, you will co-own the technical direction of the AI platform, partner closely with engineers and quantitative researchers, and report directly to our Chief AI Officer.
Your priorities in this role will thus include:
- Design and evolve the agentic core behind Aion and our AI products: agent loops and frameworks, tool use, memory and retrieval, and the skills and sidecars that agents execute.
- Build AI research systems end to end — from a fuzzy investment question through prototyping into production — encoding the investment team's expertise as context, specs, and evals rather than hard-coded pipelines.
- Give agents performant, safe access to BIT's data and capabilities, including self-serve interfaces that let the investment team create and run their own agents and workflows.
- Make probabilistic systems trustworthy: automated and human-in-the-loop evaluations, regression tests, and monitoring of accuracy, latency, and cost.
- Run LLMOps at scale: versioning of prompts, configs, and models; CI/CD; scheduling and monitoring of autonomous agent runs; fallbacks and cost control — and benchmark and select the right model per task, including open-source options.
- Stay at the frontier of LLM and agentic techniques, bring new ideas into production rather than leaving them as experiments, and act as an internal AI expert across the firm.
The experience you bring
- Systemic thinking: for anything you build, you can articulate the outcome you want to achieve, what you are constructing, how you will know it works, and the feedback loop that makes it better every week.
- Excellent software engineering: clean, well-tested Python and SQL, sound systems design, and cloud infrastructure experience (AWS strongly preferred).
- Proven experience shipping and operating production LLM and agentic AI systems — with something public you can show us: an app, agent, repo, or product.
- AI-native: you have rebuilt how you yourself work around agents (Claude Code, Codex, or open-source alternatives) — orchestrating systems that check their own outputs, not just calling LLM APIs from traditional pipelines.
- Enough statistical judgment to build honest evaluations: constructing ground-truth datasets, avoiding in-sample self-deception, and quantifying whether an output is actually good
- Genuine curiosity about business strategy, investing, and financial markets.
- Demonstrated excellence: outstanding academic record, exceptional impact in previous roles, and clear indicators of exceptional analytical ability.
Nice to have
- Experience turning unstructured text or alternative data into scores, signals, or decisions with proper evaluation.
- Hands-on experience training or fine-tuning deep neural networks, ideally transformer-based.
- Open-source contributions, public technical writing, or other engagement with the broader AI/LLM community.
2 reasons why you should not apply
- You mainly use AI tools for your own productivity and have not yet taken systems into production. This role is about building and shipping AI that the whole firm relies on, so prior production experience matters here.
- You are looking for a finished, stable AI stack to operate and maintain. Our systems evolve quickly as the frontier moves, so you will spend much of your time building and reshaping them rather than keeping a fixed setup running.
Your mindset
- You are passionate about AI and emerging technologies, and you blend technical depth with strong product judgment.
- You bet on the models: when a task can be solved agentically, you reach for a frontier model with the right context and verification before a hand-structured pipeline — and you re-test that choice as models improve.
- You take pride in making other builders fast: great abstractions, reliable systems, and tooling people love to use.
- You manage for outcomes, own problems end-to-end, and communicate clearly on progress, trade-offs, and risk — proactive, entrepreneurial, and committed to BIT Capital's vision and long-term success.
What we offer
- Work at the frontier of applied AI in asset management, with petabytes of data and the most cutting-edge tools.
- Real technical direction and ownership from day one, as the internal expert while the frontier moves.
- Contribute to a unique success story in European asset management, surrounded by the most knowledgeable experts in the industry.
- A motivated, highly international team (15 nationalities) that enjoys getting together in person in Berlin and strongly supports each other.
- Flat hierarchies and direct interaction with our management team.
- Competitive compensation, relocation, and visa support if needed.
