Machine Learning Resident - Client: Publicus (12 month term)
Core
Building deployable ML systems for multi-hop reasoning, entity resolution, and decision-making under uncertainty over a government procurement knowledge graph.
Role type
Applied Machine Learning Scientist / Research Engineer
Builds
Production-ready intelligence systems for government procurement matching and risk analysis
Domain
Government procurement, Knowledge graphs, LLMs, Sequential decision-making
Deliverable
production ML models
Required skills
Machine learning, Knowledge graphs, Entity resolution and linking, Large Language Models (LLMs), Graph-based reasoning, Sequential decision-making, Uncertainty estimation, Python, PyTorch, Hugging Face Transformers, vLLM, NetworkX, PyTorch Geometric
Preferred skills
Reinforcement learning, Bayesian decision theory, Time-to-event models, Trustworthy AI evaluation
Technologies
Python, PyTorch, Hugging Face Transformers, vLLM, NetworkX, PyTorch Geometric
Responsibilities
Design and ship multi-hop reasoning systems over procurement graphs; Build reliable entity resolution and linking mechanisms; Develop matching and competitive intelligence models; Create early-warning signals for contract risks; Design verifiable evaluation frameworks and benchmarks; Conduct applied research in graph reasoning and uncertainty estimation; Engage with client teams to deliver MVPs and reports
Seniority
Mid-Senior, hands-on IC