Research engineer building intelligent AI systems at scale
Specializing in LLM architectures, production ML, and systems that work in the real world.
Get in touchAbout
I build AI systems that bridge research and production.
Currently at Vanderbilt University's College of Connected Computing, building production GenAI systems and training the next generation of AI engineers.
Clarity over complexity
Simple systems that scale beat complex ones that break.
Production-first
Research without deployment is just theory.
Depth over breadth
Deep expertise in fewer things beats surface knowledge of many.
Scale matters
Building for 10 users vs 250K users are different problems.
Selected Work
Semantic Product Search Engine
Production MLBuilt and deployed a vector-based semantic search engine processing 250,000+ products for an enterprise e-commerce platform. Achieved 90%+ accuracy through custom embedding models and intelligent ranking algorithms.
35% increase in user engagement, 60% reduction in search latency through Redis caching architecture.
Hybrid retrieval combining semantic embeddings with traditional search. Custom ranking models tuned for domain-specific relevance. Distributed architecture with auto-scaling on AWS.
Python, FAISS, Redis, PostgreSQL, AWS Lambda, Docker
DataPulse
Best in ShowAI-powered customer segmentation platform combining Snowflake data pipelines with LLM-driven insights. Winner of "Best in Show" award for innovative application of data analytics and generative AI.
Traditional segmentation lacks context. How do you make data actionable without requiring SQL expertise?
Built natural language interface over Snowflake using Gemma LLM. Automated segment generation with AI-driven insights. Real-time analytics pipeline processing millions of customer records.
Snowflake, Gemma LLM, Python, React, ETL pipelines
AI Interview Agent
GenAIMultimodal interview simulation system combining Ollama LLM with Whisper speech recognition. Conducts natural voice conversations analyzing both content and delivery.
Real-time speech-to-text pipeline. Context-aware dialogue management. Preloaded with job descriptions and resume parsing for targeted questions.
Local LLM inference with Ollama. Whisper for speech recognition. FastAPI backend with WebSocket for real-time interaction. Modular design for easy model swapping.
Ollama, Whisper, Python, FastAPI, WebSockets
TeleMedi
Healthcare AITelemedicine platform with Llama 3.1 integration for intelligent symptom analysis and diagnosis assistance. Bridges the gap between patients and medical professionals through AI-augmented consultations.
Healthcare requires accuracy, privacy, and explainability. Can't sacrifice any for convenience.
Local LLM deployment for privacy. Structured prompting for consistent medical reasoning. Integration with video consultation platform. Audit logging for compliance.
Llama 3.1, React, Flask, PostgreSQL, WebRTC
More projects and experiments on the full projects page
Experience
Generative AI Research Engineer
Vanderbilt University College of Connected Computing
Building GenAI software systems at Vanderbilt's College of Connected Computing. Architecting LLM-powered applications, deploying AI infrastructure on AWS, and creating comprehensive training programs for engineering teams.
Data Engineer
Haystream
Architected enterprise Snowflake data pipelines and ETL workflows. Optimized query performance by 40% through warehouse tuning and intelligent caching strategies.
Generative AI Engineer
Astoria AI
Built RAG pipelines with Pinecone vector database achieving 25% accuracy improvement. Deployed real-time AI agents with sub-second response times.
Graduate Teaching Assistant
Stony Brook University
Supported course instruction and student learning in computer science graduate programs, providing academic guidance and technical assistance.
Software Engineer (Solutions & Machine Learning)
Skit
Spearheaded the development of AI-powered Voice Bots tailored for fintech clients, automating over 100,000+ customer service calls daily, streamlining operations, and improving customer experience. Collaborated cross-functionally with Delivery, Infra, and Product teams to design robust solutions for non-standard use cases. Built and optimized data pipelines and dashboards using PL/SQL and PostgreSQL in Metabase, uncovering key customer behavior insights. Developed and deployed RESTful APIs using Django/Python. Fine-tuned multilingual language models to boost voice bot accuracy, scalability, and personalization.
AI Team Lead
Eunimart
Led AI team at Eunimart, ensuring 100% project success across eight diverse projects. Implemented event-driven messaging systems using Kafka/RabbitMQ for real-time data streaming. Developed cutting-edge AI solutions to optimize business processes and drive innovation. Collaborated with cross-functional teams to deliver impactful results and drive business growth.
Software Developer
Eunimart
Developed a Similar Products Predictor using VGG16 & BERT, enhancing product recommendation accuracy. Implemented a Sales Prediction Model with XGBoost and Prophet, improving demand forecasting accuracy. Created an Image Optimizer service using OpenCV, reducing dependency on professional photo shoots.
Intern
Eunimart
Designed and deployed an automated Web Scraper (Selenium) for e-commerce data collection, improving data quality by 30%. Gathered over 50 million product records to enhance market analysis and decision-making processes.
Current Focus
Building intelligent AI systems that scale.
Exploring LLM architecture patterns for production environments. Researching efficient prompt engineering techniques for domain-specific applications. Investigating RAG pipeline optimizations for enterprise data at scale.
Currently working on frameworks for deploying GenAI systems with reliability, observability, and cost-efficiency.
Contact
Let's build something meaningful.
Nashville, Tennessee