About Me
I build real solutions, whether it’s data-driven insights or production-grade software. Experienced across AI/ML, backend engineering, and iOS development, I focus on shipping clean, scalable systems, not buzzwords. I’ve…
I build real solutions, whether it’s data-driven insights or production-grade software. Experienced across AI/ML, backend engineering, and iOS development, I focus on shipping clean, scalable systems, not buzzwords. I’ve delivered projects from forecasting models to full-stack apps, and I’m always hunting the next tough problem worth solving.
Experience
AI Backend Engineer
Slashed AI interview response latency by 85% (20 to 3 seconds) through end-to-end streaming architecture redesign, eliminating blocking operations across the speech-to-text, LLM, and text-to-speech pipeline.
Implemented real-time audio chunk processing with Google Cloud STT v1, streaming transcripts directly to Gemini 2.5 Flash for dynamic question generation with zero buffering delays.
Engineered hybrid sentence-token streaming for LLM outputs feeding into Gemini 2.5 Flash TTS, generating raw PCM chunks converted to WAV format and delivered via WebSocket for near-instantaneous audio playback.
Architected frontend audio queue system for sequential streaming playback, ensuring smooth user experience with synchronized audio delivery and eliminating choppy playback artifacts.
AI Backend Engineer
Slashed AI interview response latency, Implemented real-time audio chunk processing, Engineered hybrid sentence-token streaming, Architected frontend audio queue system
AI Backend
Optimized REST API for pre-warehouse file staging, cutting data engineers’ manual processing time by 75% through asynchronous workflow redesign.
Automated Excel-to-SQLX conversion by building a schema parser integrated with an LLM, reducing code generation time from 30 to 2 minutes per file at 75% accuracy.
Migrated chatbot infrastructure from Amazon Lex (AWS) to Google Dialogflow CX, achieving 40% higher conversation completion rates and 25% fewer escalations.
Rewrote backend Lambda functions for deployment on Google Cloud Run and transitioned the infrastructure from AWS to GCP.
Preserved 100% integration integrity during migration, including Facebook Messenger, via adapter patterns and zero-downtime testing frameworks, maintaining 99.95% uptime.
AI Backend
Optimized REST API for pre-warehouse file staging, Automated Excel-to-SQLX conversion, Migrated chatbot infrastructure, Rewrote backend Lambda functions, Preserved 100% integration integrity