AI Systems Engineer building production systems across machine learning, backend, and web.
Note
Specializing in the design and development of reliable AI-native architectures, low-latency computer vision applications, and production-ready machine learning pipelines.
• AI Agents • Computer Vision Systems • ML Pipelines • Backend Infrastructure • Full-stack Applications
| Metric | Project / Domain | Impact & Context |
|---|---|---|
| 95–100% | NL→SQL Agent | Benchmark accuracy on text-to-SQL conversions |
| <100ms | AEGIS Surveillance | Real-time computer vision inference latency |
| 30K+ SKUs | Demand Forecasting | Scalable pipeline for SKU inventory predictions |
| 3,364 Players | Scouting Database | Aggregated player performance database |
| 10.6% | RL Traffic System | Efficiency improvement in traffic routing optimization |
Real-time computer vision system designed to detect camera tampering such as blur, glare, and replay attacks.
Highlights
- Processes video frames with <100ms computer vision inference latency
- Detects camera tampering like lens blurring, high-intensity glare, and replay loops
- Features a real-time monitoring dashboard with low-latency Socket.IO communication
Links
End-to-end system that converts natural language queries into validated SQL and executes analytical workflows.
Highlights
- Achieves 95–100% NL→SQL benchmark accuracy
- Utilizes schema-aware retrieval to select optimal context for complex databases
- Integrates self-correcting validation loops to repair invalid SQL queries before execution
Links
Semantic search and graph-based exploration system for interactive content discovery with optimized query performance.
Highlights
- Powers interactive movie recommendations via multi-dimensional semantic search
- Implements Redis caching to minimize query latency for frequently requested paths
- Utilizes Pinecone vector databases for fast, scalable nearest-neighbor exploration
Links
Production website developed for a non-profit organization with focus on accessibility, performance, and responsive design.
Highlights
- Designed with 100% responsive and accessible web components (WCAG compliant)
- Optimizes page speed and visual performance for diverse device screens
- Built using modern web standards to ensure search engine optimization (SEO)
Links
- Agentic AI architectures
- Evaluation-driven LLM systems
- Production ML pipelines
- Distributed backend systems
