Job Description

KEY ACCOUNTABILITIES

● Build ML solutions for decision-making problems: planning, sequencing, routing,
allocation, and resource utilization.
● Prototype fast using agentic coding tools (e.g., Claude Code-style workflows):
generate scaffolds, refactor, write tests, iterate on experiments—while maintaining
strong engineering discipline.
● Develop and evaluate models in areas like:
○ Optimization & solvers: MILP/CP-SAT, heuristics/metaheuristics, constraint
programming, search methods
○ Deep RL / Decision Intelligence: RL baselines, offline RL, bandits,
MCTS-style planning, policy/value learning
○ Predictive ML: forecasting and estimation models that feed decision systems
● Design robust evaluation harnesses: offline simulation, counterfactual testing,
ablations, and scenario analysis; define KPIs and acceptance thresholds.
● Collaborate with ML engineers to support productionization: latency/throughput
constraints, monitoring, reproducibility, model versioning, and safe rollout.
● Write clear technical documentation and communicate findings to both technical and
non-technical stakeholders.

What We’re Looking For (Required)
● 0–5 years experience in applied ML / data science / applied research (internships,
thesis work, and strong project portfolios count).
● Demonstrated experience using agentic coding assistants in real development
(e.g., Claude Code, similar agentic coding environments) to accelerate
iteration—without sacrificing code quality.
● Strong Python skills and comfort with ML tooling (PyTorch preferred; TensorFlow ok).
● Solid foundations in algorithms, probability/statistics, and experimental design.
● Ability to translate messy real-world problems into clear formulations and measurable
success metrics.

Strong Plus / Preferred
● Prior work in Deep RL (a strong differentiator), such as:
○ PPO/SAC/DQN style methods, offline RL, imitation learning, MCTS/planning
hybrids
○ Building environments/simulators, reward design, stability/debugging,
evaluation
● Experience with simulation-based evaluation or digital twins (even lightweight
simulators).
● Familiarity with MLOps basics: MLflow, Docker, CI/CD, model monitoring.
● Domain exposure to logistics/supply chain/industrial operations (nice-to-have, not
required).

Tools & Tech (Indicative)
Python, PyTorch, OR-Tools / solver stacks, RL libraries (Ray RLlib / Stable Baselines), SQL,
Docker, Git, MLflow; cloud platforms a plus.

#LI-MP1