Vetted Reinforcement Learning Professionals

“Cornell master’s student (CS/Stats) focused on research-heavy ML projects: implemented a sparsity-driven RL approach (DAPD + Soft Actor-Critic) that maintained stable learning even with ~95% of weights removed in OpenAI Gym continuous-control tasks. Also worked on diffusion-based computer vision with conditioning and latency-focused U-Net choices, and scaled unsupervised community detection on a 50k-node/800k-edge Reddit graph via BFS subgraph sampling.”

“Full-stack engineer with startup ownership experience (Aiir) building 15+ TypeScript/Go microservice APIs on GCP Cloud Run with Kafka-based async event streaming and React CRM integrations for billing/analytics. Strong post-launch operator who tuned Oracle performance (partitioning/indexing/query optimization) and validated a 23% retrieval-time reduction via AWR, and has a quality/DevSecOps mindset (94% Pytest coverage, GitHub Actions, SonarQube, Twistlock, CloudWatch) including migrating 18+ production CI/CD pipelines.”

“Robotics software engineer with hands-on experience across reinforcement learning and ROS/ROS2, including a project teaching Boston Dynamics Spot to open a door by combining vision-based pose estimation with SAC-trained IK and a walking policy in MuJoCo. Previously built ROS Noetic control for surgical robots using RCM with MoveIt IK and achieved sub-0.02s latency via threading; also participated in a NASA ROS2 space simulation building rover teleop and sensor-driven mapping.”

“Robotics software engineer who built an end-to-end mobile manipulation platform (Franka Panda on a Clearpath Ridgeback) for a simulated-kitchen human-robot interaction study with natural speech commands, implemented in Python/ROS. Has hands-on experience integrating diverse sensors (RealSense, LiDAR, biosignals) with deep learning frameworks (PyTorch, Hugging Face) and fine-tuning GPT-Neo, plus simulation (Gazebo) and modern deployment practices (Docker/Kubernetes, CI/CD).”

“Robotics software engineer who owned the end-to-end navigation stack for a mobile manipulation robot (Cone-E), integrating ZED-2i SLAM into a real-time occupancy grid with live obstacle avoidance, A* planning, and lookahead control. Strong in real-time debugging and stability improvements (goal snapping/locking, obstacle persistence, rate-limited replanning) and validates changes on hardware, supported by simulation (Gazebo/Webots) and Docker/CI-based testing.”

“Early-career robotics engineer who has integrated LLM/NLP command interfaces (OpenAI/LLaMA) into ROS-controlled industrial manipulators and built data-driven controls for underwater soft robotic actuators. Combines hands-on fabrication (balloon actuator with embedded copper traces) with sensor debugging (IMU/Aurora) and simulation work in Gazebo, with practical exposure to edge deployment constraints on Jetson Nano and model quantization.”

“Head of Robotics at Orbital Composites who built the full ROS-based software stack ("OrbOS") for multi-robot KUKA-arm 3D printing of carbon-fiber composite geometries. Deep in motion planning and real-time control—created a pipeline to turn underconstrained toolpaths into optimal joint trajectories at massive command scale, plus redundant safety gatekeeping and a 250Hz synchronized robot+extruder controller. Has operated in a technical-founder capacity for ~6 years and values an engineering culture of openness, honesty, accountability, and high ownership.”