Hawk - Senior Full Stack Engineer (Agentic AI)
Requirements
• 4+ years of professional software engineering experience, with substantial hands-on time in both Python and TypeScript/React. (Note: We are open to exceptional candidates with 2-3 years of experience if they demonstrate a remarkably strong track record.) • Deep backend expertise: Python, FastAPI, Postgres, Redis, and comfortable operating services on Kubernetes in production. You understand the difference between code that works and code that works at 3am under load. • Strong frontend chops: React, TypeScript, and modern component libraries (shadcn/ui or Material UI). You don't need to be a designer, but you care about how the thing looks and feels and you can build a clean, performant UI without supervision. • Genuinely full-stack mindset. You don't think of yourself as "a backend engineer who can do some frontend" or vice versa — you think of features, and you build them. • Hands-on fluency with agentic coding tools (Claude Code, Cursor, or equivalent) as part of your daily workflow. You have an opinion about what they're good at, what they're not, and how to get the most out of them. • Excellent holistic thinking and problem-solving skills. You shape the product rather than execute pre-defined tickets. • Excellent communicator with a collaborative mindset, capable of articulating technical trade-offs to both technical and non-technical stakeholders. • Production experience with LangGraph (or equivalent agent orchestration frameworks) and Langfuse (or equivalent LLM observability/eval tooling). • Prior experience in Anti-Financial Crime (AML, fraud, compliance, or investigations tooling) — or a deep curiosity to go learn the domain fast.
Responsibilities
• Design, build, and operate the backend systems that power the Investigation Agent: LangGraph workflows, FastAPI services, Postgres and Redis data layers, Kubernetes deployments, and Langfuse-based observability and evaluation. • Build the frontend surface that makes the agent's work legible: React + TypeScript components in shadcn/ui that surface agent reasoning, evidence, and explainability so analysts can trust, challenge, or take over. • Own features end-to-end. You take a problem from spec to production — backend, frontend, infra, observability, evals — without waiting for someone else to finish their part first. • Make the agent reliable. Build the eval harnesses, tracing, and feedback loops in Langfuse that turn "the model usually does the right thing" into "the model demonstrably does the right thing, and we know when it doesn't." • Partner closely with Product/UX, Data Science, and the rest of Engineering to shape what gets built — not just how. You question the "why" behind features and push back when something doesn't feel right. • Champion an AI-native engineering workflow. You use Claude Code, Cursor, or equivalent agentic coding tools as a daily part of how you work, and you help raise the bar for how the team uses them. If you find yourself doing the same thing twice, your natural instinct is to automate it. • Help establish engineering best practices for agentic products at Hawk: how we test agents, how we deploy them, how we monitor them, how we roll back when they misbehave.
Benefits
• Please enter the annualized pay in € • Do you agree to allow Hawk to contact you about job opportunities for up to 1 year? • Hawk's Recruiting Privacy Policy • Note: The consent period lasts for 1 year • Submit Application
Apply in one click
Upload My Resume
Drop here or click to browse · Tap to choose · PDF, DOCX, DOC, RTF, TXT