I build dependable autonomy software for industrial and mobile robots. My work spans perception, localization, sensor fusion, navigation, motion planning, simulation, fleet observability, and accelerated edge inference—connecting C++ and Python systems to real production infrastructure.
| Area | Technologies and responsibilities |
|---|---|
| Autonomy | ROS2, Nav2, TF2, SLAM, recovery behavior, mission state, DDS/Fast DDS |
| Perception and localization | LiDAR, stereo/RGB-D cameras, IMU, wheel encoders, OpenCV, PCL, EKF/UKF |
| Planning and control | A*, Dijkstra, RRT*, MPC, PID, MoveIt2, ROS2 Control, trajectory execution |
| Edge AI | NVIDIA Jetson, CUDA, TensorRT, ONNX, PyTorch, latency and CPU optimization |
| Production systems | C++17/20, Python, AWS, Docker, CMake, Linux, CI/CD, MCAP, telemetry and incident replay |
- Supported autonomy services for 300+ machines executing 5,000+ daily missions.
- Built telemetry workflows processing 10M+ robot events per day.
- Reduced recurring autonomy-investigation time from approximately 4 hours to under 90 minutes.
- Reduced unnecessary robot stops by 18% and CPU consumption by 24%.
- Brought critical edge-perception latency to approximately 30 ms under concurrent workloads.
- Expanded validation to 15,000+ simulated scenarios, reducing physical test time by nearly 35%.
| Repository | What it demonstrates | Verification |
|---|---|---|
| Aegis Runtime Assurance | Probabilistic collision prediction, Simplex-style command supervision, deterministic safe-control search, and counterfactual decision evidence | Live safety case · 7 tests · 6.75 ms p95 |
| NovaForge Sim | Deterministic multi-robot simulation with scenario acceptance gates, sensor modeling, planner experiments, and regression evidence | Live simulator · tests + CI |
| Autonomy Log Triage | Deterministic incident detection, severity classification, and heartbeat-gap analysis for synthetic robot events | Unit tests + CI |
| EKF Sensor Fusion Lab | Transparent 2D state estimation, asynchronous updates, Joseph-form covariance, and seeded simulation | Unit tests + CI |
| Robot Fleet Observability | Explainable fleet-health scoring and a dependency-free read-only telemetry API | Unit tests + CI |
These repositories use synthetic data and independent implementations. They contain no employer source code, operational data, maps, facility details, or proprietary interfaces.
- EKF Sensor Fusion Lab walkthrough — model assumptions, equations, implementation, numerical safeguards, tests, and the reproducible RMSE result.
- Autonomy Log Triage walkthrough — event schema, validation, incident rules, heartbeat-gap calculation, deterministic reporting, and system boundaries.
Each walkthrough is paired with an evidence record so reviewers can reproduce the public result and distinguish implemented capability from the production roadmap.
The public Robotics Open Source Roadmap tracks ROS 2 integration, control-barrier-function research, uncertainty calibration, hardware-in-the-loop validation, simulation extensions, and contributor-ready issues across these projects.
- Agentic RAG with LangGraph — retrieval grading, query rewriting, and self-correcting control flow.
- Speech Emotion Detection — MFCC-based speech classification with CNN and LSTM models.
- Freight Document Agent — traceable extraction of structured freight information.
- Robotics Software Engineer, Amazon Robotics — Apr 2025–Present
- Robotics Software Engineer, Siemens — May 2021–Dec 2023
- M.S., Computer Science — Roosevelt University
- B.E., Computer Engineering — Saffrony Institute of Technology
Professional certificates
- Machine Learning Operations (MLOps) — DeepLearning.AI / Coursera
- Introduction to Generative AI — Google Cloud
- IBM AI Developer Professional Certificate — IBM
- Google Data Analytics Professional Certificate — Google
Selected coursework
- Deep Learning, Intelligent Systems, Data Mining, and Big Data
- Computing with Data in Python, Advanced Algorithms, and Systems Programming Concepts
- Advanced Computer Architecture and Cryptography
I am interested in robotics roles where autonomy must remain safe, observable, computationally disciplined, and recoverable outside controlled demonstrations—especially runtime assurance, robot fleets, mobile manipulation, perception, localization, navigation, and accelerated edge inference.
C++17/20 Python ROS2 Nav2 DDS/Fast DDS TF2 Behavior Trees MoveIt2 ROS2 Control OpenCV PCL SLAM EKF/UKF CUDA TensorRT ONNX Gazebo Isaac Sim HIL/SIL AWS Docker CMake Linux MCAP Jenkins GitLab CI Real-Time Systems
Open to discussing robotics software, autonomous systems, simulation, and production reliability opportunities in the United States.