Machine Learning Engineer

Lovelace AI•Pittsburgh, PA

About The Position

As a Machine Learning Engineer at Lovelace, you will own machine-learning problems end to end, from framing ambiguous questions to shipping production systems. The role has no narrow scope or ticket queue; engineers work independently across the codebase, deciding what data matters, which models are appropriate, and how to build the surrounding systems. It is roughly half research and half engineering, on a team that reports to the Chief Scientist. Engineers at Lovelace work alongside AI coding agents, and the role emphasizes judgment: deciding what to build, identifying errors in agent output, and owning design decisions. Industry experience is not required; strong research-lab backgrounds are welcome. The company seeks people who can ramp up quickly, own difficult problems, and think critically about their own work and the work of AI agents.

Requirements

Track record of owning ambiguous problems end to end in a research lab, startup, or personal projects.
Strong machine-learning fundamentals, including probabilistic foundations, modern deep learning, and LLM architectures.
Experience building the systems around models, including data pipelines, evaluation, deployment, and operation at scale.
Hands-on experience with AI coding agents and a clear understanding of where they can go wrong.
Expert-level Python and a strong testing discipline; understands how their code works.
Confidence to challenge requirements, teammates, and AI agents with evidence.

Nice To Haves

Experience with NLP, knowledge graphs, entity resolution, or agentic systems.
Go or other systems-programming experience.
Open-source contributions, side projects, or publications with working code.
Background at a strong research lab; a PhD is welcome but not required.

Responsibilities

Frame ambiguous product needs as precise machine-learning problems and set requirements.
Determine useful data, suitable model types, and when approaches beyond fine-tuning are needed.
Direct AI coding agents, critically review their output, and challenge incorrect design decisions.
Design systems around models, including data pipelines, serving, evaluation, and monitoring.
Choose metrics that reflect real outcomes and identify when metrics are misleading.
Stand up new jobs, make changes across the codebase, and ship NLP and LLM systems to production while owning their correctness.
Challenge assumptions, including those of teammates and oneself, to reach the right answer.