ποΈ AI Engineer | Speech Recognition (ASR) | Real-Time AI Agents | MCP
I build production-grade Speech Recognition systems and real-time AI agents with a strong focus on low latency, streaming inference, and scalable architectures. My work spans decoding algorithms, agent orchestration, and real-time AI pipelines.
- ποΈ Building Speech Recognition (ASR) systems (training, decoding, evaluation)
- β‘ Real-time & streaming inference systems
- π€ Designing AI Agents using MCP & multi-agent architectures
- π§© Deep focus on decoding strategies (Greedy, Beam Search)
- ποΈ Production-ready AI & backend system design
- PyTorch
- ONNX / ONNX Runtime
- Streaming ASR (CTC, Transducer)
- Beam Search & Greedy Decoding
- WebSockets
- gRPC
- Low-latency streaming pipelines
- MCP (Model Context Protocol)
- Multi-Agent Systems
- Tool-calling & orchestration
- Realtime agent frameworks
- Python, Node.js , Java , C++ , C , Dart
- FastAPI, Flask , Django , SpringBoot
- Docker, Nginx ,
- Linux
- ποΈ Building ASR systems for indian languages and global languages
- β‘ Fixing beam-search hallucinations in silent audio
- π€ Real-time AI agents using MCP
- π§ Multi-agent coordination & tool execution
- π Deploying low-latency AI services
