Speech AI / Machine Learning Engineer focused on building real-world voice and conversational systems.
I work on end-to-end pipelines — from speech recognition and voice interfaces to scalable backend infrastructure powering AI applications.
Building voice-first systems for real-world environments where typing is not the interface.
- Speech-to-Text systems (ASR, Whisper fine-tuning, VAD)
- Voice-first applications for multilingual users
- Real-time AI pipelines (streaming, async workers, queues)
- LLM-powered systems (RAG, agents, conversational workflows)
- Scalable backend systems (FastAPI, microservices, cloud deployment)
Languages
Technologies / Frameworks
ML / LLM
Deployment
- Improved ASR performance for medical speech (WER ↓ significantly)
- Built production-ready async pipelines handling real-time audio workloads
- Experience deploying AI systems used in real environments
- 🥇 Winner — National-level BharatGen AI Hackathon for developing JanVaani, an AI voice bot that automates grievance form submission
- 📊 Top 3 — Machine Learning Kaggle Competition
- 🥈 Runner-up — Hacksprint Hackathon for developing a sustainability-focused web application
- Hosted, directed, and wrote a mime play at a college cultural event attended by 2,000+ members
- Email: utsav.ashutosh@gmail.com
- LinkedIn: linkedin.com/in/ashutosh-utsav
- GitHub: github.com/ashutosh-utsav