Engineer. Builder. Not a prompt writer.
I design and ship systems at the intersection of AI, automation, and real-time engineering. My work focuses on making AI components actually useful in production — reliable pipelines, not demos.
What I spend most of my time on:
→ Designing AI agent architectures that handle real workflows
→ Building computer vision pipelines for detection and recognition tasks
→ Wiring backend systems — async APIs, event queues, orchestration layers
→ Shipping multimodal AI applications (speech + vision + language combined)
→ Automating operational processes with tools like n8n + custom FastAPI services
Multimodal AI · Speech Analysis · Vision · Evaluation Pipeline
An end-to-end interview evaluation engine that processes audio, video, and language simultaneously. Analyzes speech clarity, confidence signals from facial tracking, and response quality — outputs structured feedback across multiple dimensions.
Python FastAPI OpenCV Whisper Multimodal AI Real-time Processing
Android · Bluetooth Automation · Real-Time Triggers · Background Services
Android automation application built around Bluetooth device events as real-time triggers. Designed for background reliability — stays alive through aggressive battery management and handles edge cases other automation tools miss.
Kotlin Android Bluetooth APIs Background Services Event-Driven Architecture
FastAPI · React · Google Maps API · Real-Time Prediction
Full-stack traffic prediction platform combining live API data with ML forecasting. React dashboard surfaces real-time predictions with route-level insights. Backend orchestrates multiple data feeds with async processing.
Python FastAPI React Google Maps API Real-Time Data REST APIs
Computer Vision · Gesture Recognition · OpenCV
Real-time gesture classification pipeline for sign language recognition. Custom computer vision model trained on hand landmark data. Optimized for low-latency inference on consumer hardware.
Python OpenCV Computer Vision MediaPipe ML Inference
OpenCV · Image/Video Analysis · Forensic AI
Detection pipeline for identifying AI-generated and manipulated media. Works on both static images and video frames. Built with a focus on speed and reliability across varied compression artifacts.
Python OpenCV Image Analysis Video Processing Detection Models
