Technical AI Product Manager

About the Role

As a Technical AI Product Manager, you will oversee AI-driven product initiatives within legal workflows, bridging the gap between product management and technical execution. You identify valuable AI opportunities, validate concepts, and lead successful deployments into production. This role combines a strong product background with deep technical knowledge rather than pure engineering.

Key Responsibilities

• Discover high-impact AI applications in legal processes, especially in areas without dedicated squads.
• Validate ideas swiftly by engaging with lawyers, legal operations, and insurers; quickly prototype promising solutions.
• Lead the hands-on implementation of winning features, including routing, chaining, summarization, and autonomous agent loops.
• Once mature, hand over features to squads with comprehensive documentation, evaluation plans, and transition strategies.
• Serve as a sparring partner for squads, advising on feasibility, cost, model selection, and risk management for AI features.
• Design and assist with evaluations, golden datasets, and guardrails, acting as the primary point of contact.
• Review concepts before roadmap inclusion, addressing prompt-injection and compliance risks such as GDPR and legal statutes.
• Share AI expertise in the product team, fostering best practices like prompt-engineering standards, LLM observability, and autonomous coding workflows.

Qualifications

• Strong background in both product management and engineering, with demonstrated end-to-end project ownership.
• Proven ability to develop tangible products from ambiguous challenges.
• Several years of experience working extensively with AI, deploying multiple LLM systems to real users.
• Proficiency in Python and consistent use of tools like Git, testing, CI/CD, and code reviews.
• In-depth knowledge of LangGraph, LangChain, or similar technologies, including underlying mechanics such as tool calling, structured outputs, and streaming. Comfortable with LLM provider APIs.
• Solid experience preparing evals, golden datasets, and guardrails prior to releases; adept in applying appropriate prompt engineering techniques.
• Fluent German (minimum C1 level) to navigate the legal domain, alongside business-level English proficiency.

Soft Skills

• Highly self-driven with the ability to independently prioritize work and say no when necessary.
• Effective communicator bridging technical trade-offs to legal stakeholders and conversely translating legal workflows for engineering teams.

Work Environment and Benefits

• Office-first culture with 3 days onsite in a modern, daylight-filled, air-conditioned Berlin office near Gleisdreieck and Potsdamer Platz, plus 2 days remote flexibility.
• Full coverage of the Germany public transport ticket.
• Company pension scheme with 20% employer contribution.
• Access to a corporate benefits platform offering discounts on various brands.
• Dog-friendly office environment welcoming four-legged team members.
• Regular company and team events including after-work fun featuring Switch, PS5, darts, and foosball.
• Complimentary beverages, fresh fruit, and muesli.

Culture and Collaboration

• Agile mindset with open communication valuing every voice.
• Feedback-oriented culture focusing on shared growth.

Recruitment Process

• Submit a concise application as a Senior AI Product Engineer.
• Participate in a ~20-minute introductory conversation with HR.
• Engage in an approximately 60-minute onsite interview with the hiring manager.
• Attend a ~30-minute discussion with management.
• Receive a timely job offer.

Additional Information

Legalhero is a modern platform transforming legal protection insurance and attorney services by integrating directly with insurers' systems. Founded in 2017 and based in Berlin, we help hundreds of thousands of policyholders with the aid of cutting-edge AI technologies. We champion an inclusive, equal-opportunity, and non-discriminatory workplace. If you have accessibility needs during the recruitment process, please let us know.