Full Stack Engineer at Consensus AI

Implemented ETL pipeline using Python and Apache Spark; built supplement and herb database based on PubMed medical research data. Optimized LLaMA-38B model to reduce size by 75% and latency by 50%, maintained 0.95 recall and 30 BLEU score. Developed a RAG-enabled medical research chatbot using LangChain with few-shot and chain-of-thought prompting. Built asynchronous backend with FastAPI and Redis caching; migrated backend services to Go microservices on AWS Lambda for improved scalability. Implemented event-driven architecture using AWS Services, Apache Kafka and RabbitMQ for real-time inventory updates across WhatsApp and Twilio, reducing product availability discrepancies by 25%. Led CI/CD adoption with AWS CDK, Jenkins, and GitHub Actions, cutting deployment times and improving reliability.

AI Engineer at Air AI

Developed a personalized product recommendation system leveraging transformer-based models (BERT, T5) and PyTorch, contributing to a 32% conversion lift and $2.5M in annual revenue. Integrated ElevenLabs TTS for audio personalization; built Kaldi-based ASR; designed a WebSocket-based robust notification system handling over 1.5M real-time notifications weekly. Implemented OAuth 2.0 and strengthened security, reducing security incidents to zero for over 12 months.

Full Stack Engineer at Consensus AI

Implemented ETL pipeline using Python and Apache Spark and built supplement and herb database based on PubMed medical research data. Optimized LLaMA-38B models by 75% in size and 50% in latency, while maintaining 0.95 recall and 30 BLEU score. Developed a RAG-enabled medical research chatbot using LangChain with few-shot and chain-of-thought prompting, improving lead qualification by 40% and content personalization by 25%. Built an asynchronous backend with FastAPI and Redis caching. Refactored backend with Go microservices on AWS Lambda and Docker, reducing API response times from 1.2s to 0.7s and enabling scalability for 200k+ concurrent users. Implemented event-driven architecture using AWS services, Apache Kafka, and RabbitMQ for real-time inventory updates across WhatsApp and Twilio, reducing product availability discrepancies by 25%. Led CI/CD adoption with AWS CDK, Jenkins, and GitHub Actions, cutting deployment times and ensuring 100% deployment success.

AI Engineer at Air AI

Developed a personalized product recommendation system leveraging transformer-based models (BERT, T5) and PyTorch for an e-commerce client, increasing conversions by 32% and generating $2.5 million in annual revenue. Integrated ElevenLabs TTS for audio personalization. Built ASR with Kaldi components. Designed a robust notification system using WebSockets with Socket.IO, handling over 1.5 million real-time notifications weekly and boosting user engagement. Implemented OAuth 2.0 and security protocols, reducing security incidents to zero for 12+ consecutive months.

Full Stack Engineer at Consensus AI

Led backend and data engineering initiatives, building an ETL pipeline with Python and Apache Spark to curate PubMed-derived supplement and herb data. Optimized LLaMA-3 8B models by 75% in size and 50% in latency while preserving 0.95 recall and 30 BLEU score. Created a RAG-enabled medical research chatbot with LangChain using few-shot and chain-of-thought prompting, boosting lead qualification and content personalization. Built an asynchronous FastAPI backend with Redis caching and deployed Go microservices on AWS Lambda and Docker to cut API response times from 1.2s to 0.7s. Implemented event-driven inventory updates across WhatsApp and Twilio using AWS, Kafka, and RabbitMQ, and led CI/CD adoption with AWS CDK, Jenkins, and GitHub Actions to reduce deployment time.