Note: The job is a remote job and is open to candidates in USA. DeepRec.ai is hiring a Senior Machine Learning Engineer specializing in reinforcement learning to build and deploy scalable decision-making agents. The role focuses on translating RL research into low-latency production systems, including algorithm scaling, simulator and reward design, real-time deployment, and formulation of business problems as Markov Decision Processes.

Responsibilities

• Algorithm & Pipeline Scaling: Implement and tune model-free, model-based, or offline RL algorithms (PPO, SAC, IMPALA, DPO/GRPO) across multi-GPU environments
• Environments & Reward Design: Build custom simulators and reward structures using Gymnasium, MuJoCo, or distributed frameworks
• Production Deployment: Package, serve, and monitor real-time RL models in high-throughput pipelines
• Problem Formulation: Map complex business problems into clear Markov Decision Processes (MDPs) alongside cross-functional teams

Skills

• • RL Expertise: 3+ years implementing RL algorithms in PyTorch or JAX with tools like Ray/RLlib, CleanRL, or TorchRL
• • Math & Fundamentals: Solid grounding in MDPs, dynamic programming, linear algebra, and probability
• • Distributed Engineering: Hands-on experience with CUDA, distributed training (DDP/vLLM), Docker, and Kubernetes
• • Software Mastery: Strong Python, OOP, CI/CD, and production system design
• • Published papers (NeurIPS, ICML, ICLR) or top finishes in RL competitions
• • Direct experience with Offline RL, Imitation Learning, or LLM alignment (RLHF/DPO)
• • Familiarity with NVIDIA Isaac Gym or Unity ML-Agents

Company Overview