Looking for a job

Manish Reddy

@manishreddy4

Message

I build production machine learning systems for multilingual speech, agentic RAG, and real-time 3D avatars.

India

Message

What I'm looking for

I'm looking to build production AI systems spanning multilingual speech, agentic RAG, generative models, and real-time avatar experiences, from experimentation through deployment and measurable evaluation.

At Flam, I build production systems for real-time conversational and 3D avatars, taking models from research through inference optimization, evaluation, and deployment.

I engineered a proprietary multilingual text-to-speech stack from scratch, including the text front-end, acoustic model, and vocoder. It serves 60+ languages, including all 22 official Indian languages, at sub-300ms time-to-first-audio.

I'm building an agentic RAG framework that decomposes multi-step requests into tool calls and executes actions rather than only describing them. I also built a self-hosted cross-lingual voice cloning pipeline and multilingual text normalization across 60+ languages.

Previously, I built a production NLP chatbot at Levyug that improved intent accuracy by 50%, reduced support-agent workload by 40%, and cut response latency by 75%. At HEEDDATA, I automated AWS infrastructure and CI/CD workflows, reducing provisioning effort by 65% and infrastructure costs by up to 30%.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer

Current

Flam

Mar 2026 - Present (5 months)

Owns model development for real-time conversational and 3D avatar systems, from research through production deployment. Engineered a proprietary multilingual TTS stack serving 60+ languages at sub-300ms latency and is building an agentic RAG framework with action execution.