AI/ML Engineer specializing in high-performance GPU systems.
I build production-ready intelligence at enterprise scale, focusing on LLMs and MLOps.
Featured Projects
sentry-SAM
Source ↗Real-time surveillance system integrating Segment Anything Model (SAM) for advanced object tracking and security monitoring.
retractly
Source ↗AI-powered document redaction tool designed for privacy-conscious data processing and sensitive information masking.
mcp-ocr
Source ↗Enterprise-grade OCR server using Model Context Protocol (MCP), enabling seamless text extraction from multi-source media.
white-board-camera
Source ↗Intelligent image processing for whiteboard captures, optimizing contrast and perspective for digital archiving.
surface-crack-detection
Source ↗Deep learning model for structural health monitoring, automatically detecting and localizing surface cracks in infrastructure.
SentinelPhishFeed
Source ↗Autonomous threat-intelligence agent that crawls and publishes 700K+ IOCs on a 3-hour refresh cycle for active phishing prevention.
Experience
- Built an automated logo annotation system leveraging linear algebra, saving 10,000+ hours/year.
- Migrated AWS GPU inference to TensorRT, cutting costs by 66% while boosting throughput by 10%.
- Developed an active ML algorithm achieving 99% precision for threat detection.
- Developed Canvas AI platform (Langgraph, FastAPI), increasing client engagement by 25%.
- Integrated TensorFlow predictive models, increasing performance by 30%.
- Designed and deployed scalable AI architectures on Azure.
Expertise
ML / Deep Learning
PyTorch, TensorFlow, Transformers, LLMs, Hugging Face, QLoRA, PEFT.
AI Infrastructure
NVIDIA CUDA, TensorRT, GPU Optimization, A100/T4, HPC.
MLOps & Systems
MLflow, Airflow, CI/CD, FastAPI, Rust, MongoDB, Azure, AWS.
Frameworks
LangChain, LangGraph, Pydantic AI, Next.js, MATLAB.