M.S. Computer Science @ University at Buffalo
Building reliable software for robots, distributed systems, and intelligent applications.
class Shubham {
public:
std::string focus = "Robotics + Systems + AI";
std::vector<std::string> languages = {
"C++", "Python", "Java", "TypeScript", "SQL"
};
std::vector<std::string> interests = {
"Robotics Software",
"Distributed Systems",
"AI Infrastructure",
"Developer Tools"
};
std::string currentlyBuilding() {
return "Reliable software for intelligent systems";
}
};- 🤖 Building ROS 2 / C++ / Python software for robotics and intelligent systems
- ⚡ Interested in robotics infrastructure, autonomy, perception & distributed systems
- 🧠 Optimized ML inference from ~92 ms → ~37 ms using ONNX + TensorRT
- ⚙️ Built backend systems supporting millions of daily transactions
- 🚀 Improved backend throughput by ~90% through performance optimization
- 🔬 Experience with sensor integration, diagnostics, telemetry & robotic validation
- 🌍 Active contributor to open-source C++ and Python projects
ROS 2 · rclcpp · TF2 · URDF · DDS · Gazebo · Isaac Sim · CAN · Serial · Ethernet · Sensor Integration
REST APIs · Microservices · Caching · Concurrency · Idempotency · CI/CD · Observability
Computer Vision · Vision Transformers · TensorRT · LLMs · RAG · LangGraph · Pinecone · AI Agents
| Area | Impact |
|---|---|
| 🤖 Robotics | Built ROS 2 perception, sensor integration, diagnostics & telemetry software |
| ⚡ ML Performance | Reduced inference latency from ~92 ms → ~37 ms |
| 🧠 Computer Vision | Improved benchmark F1 from 0.29 → 0.55 |
| ⚙️ Distributed Systems | Worked on backend systems processing millions of daily transactions |
| 🚀 Performance | Improved backend throughput by ~90% |
| 🛡️ Reliability | Implemented retries, idempotency, caching, logging & observability |
| 🧪 Validation | Conducted evaluation across 1,000+ robotic runs |
Production-grade distributed rate limiting infrastructure using Redis + Lua atomic execution.
Highlights
- ⚡ Atomic Redis + Lua rate enforcement
- 🔄 Multiple rate-limiting algorithms
- 📊 Prometheus/Grafana observability
- 🧪 Benchmark and performance suite
- 🐳 Containerized deployment
Tech: Java Redis Lua Docker Prometheus Grafana
Distributed task-processing infrastructure focused on concurrency, reliability and fault tolerance.
Focus: Java · Distributed Systems · Concurrency · Fault Tolerance
AI-powered developer tooling and workflow automation.
Focus: Python · AI Agents · Developer Tools · Automation
Generative AI experimentation with a deployment-oriented inference workflow.
Focus: Python · Generative AI · ML Inference · Jupyter
I actively contribute to external open-source projects across C++, Python, Java, developer tooling, and infrastructure — working on bug fixes, regression tests, reliability improvements, and production codebases.
C++ · Debugging · Regression Testing
- Fixed null-parent handling in orphan
ItemandTransitionrange queries - Hardened composition ancestry traversal against invalid parent access
- Added C++ regression tests covering orphan range-query paths
- Validated changes through build, tests, and CI
Contributed across projects in ecosystems including:
Microsoft · aio-libs · Meilisearch · Academy Software Foundation · Java · Python · C++
Recent contributions include:
aio-libs/aiomysqlmicrosoft/ebpf-for-windowsmeilisearch/meilisearch-javamid0/midoeko/gocachedancormier/openwebui-claude-agent-pipeyunaremaia/gfiyunaremaia/taintrace
Intelligent Systems
│
┌────────────────┼────────────────┐
│ │ │
▼ ▼ ▼
Robotics Systems AI
│ │ │
ROS 2 / C++ Distributed Computer Vision
Perception Systems ML Inference
Sensors Reliability Agentic AI
Diagnostics Performance RAG / LLMs
Simulation Infrastructure AI Infrastructure
Understand the problem
│
▼
Build the simplest reliable solution
│
▼
Measure real behavior
│
▼
Find the bottleneck
│
▼
Optimize
│
▼
Test failure cases
│
▼
Ship → Observe → Improve → Repeat
