Lexsi Labs is one of the leading frontier labs focusing on building aligned, interpretable, and safe Superintelligence. Most of our work involves creating new methodologies for efficient alignment, interpretability-led strategies, and foundational model research across enterprise and tabular data. Our mission is to build AI systems that are powerful, transparent, and trustworthy by design.
Our team thrives on deep technical rigor, hands-on ownership, and a strong bias toward building systems that actually work in production. At Lexsi.ai, we operate with a flat structure, fast feedback loops, and an expectation that everyone deeply understands what they ship.
As a
Product Manager
at Lexsi.ai, you will own the end-to-end lifecycle of ML-driven platform features. This is not a coordination role. You will be responsible for problem definition, technical clarity, correctness, evaluation, and quality. You will work closely with AI Researchers, AI engineers, SDEs, and QA to ensure every feature behaves as intended, is testable effectively, and delivers real value to users.
We are explicitly looking for someone with a
strong AI/ML background
who has grown into product ownership, not a traditional PM learning ML on the job.
Responsibilities
- Own product requirements and execution for AI/ML and platform-heavy features from problem definition to production rollout.
- Translate complex AI behavior into precise, testable product specifications and acceptance criteria.
- Work deeply with AI Researchers to collect the details of the components created and AI engineers and SDEs to ensure implementations match intended algorithms, assumptions, and constraints.
- Define what “correct” means for each feature, including metrics, evaluation logic, edge cases, and failure modes.
- Work with QA and testing teams to design meaningful test strategies that reflect real ML behavior, not just happy paths.
Requirements
- Prior experience as an ML engineer, data scientist, or research engineer.
- Experience building developer platforms, ML tooling, or infra-heavy products.
- Strong foundation in Deep Learning, LLMs, machine learning or data science with hands-on experience building or deploying AI systems.
- Familiarity with testing strategies for ML systems, including offline evaluation, simulation, and monitoring.
Nice to Have
- Exposure to interpretability, alignment, model observability, or AI safety work.
- Startup experience where product ambiguity and technical depth coexist.
What Success Looks Like
- Features ship with clear intent, correct behavior, and well-defined evaluation criteria.
- QA and testing teams understand what they are testing and why it matters.
- Engineering rework drops because specs capture real constraints upfront.
- Product velocity increases without sacrificing correctness or trust.
