Skip to content
~/avatar
Uday Devnani
actively writing

Uday Devnani

AI/ML Engineer

I build production-grade AI systems using Python, FastAPI, and modern ML frameworks. Specialized in speech-to-text pipelines, LLM-powered RAG applications, and scalable asynchronous AI workflows deployed on cloud platforms.

9 posts
34 topics
2+ years

What I Do

GenAI & LLMs

Building robust LLM applications, rigorous prompt engineering, and fine-tuning transformer models for specialized production use cases.

RAG Systems

Vector databases, semantic search, and highly-optimized retrieval pipelines.

Speech Intelligence

Scalable STT pipelines, advanced audio enhancement techniques, and speaker diarization systems engineered for high-throughput.

Experience

Backend AI/ML Developer

Jan 2024 – Present

Webs Optimization Software Solution

  • Designed AI-driven backend services with FastAPI, supporting robust, high-volume async request processing architectures.
  • Engineered end-to-end STT pipelines utilizing Whisper and Deepgram, dramatically reducing manual review bottlenecks by over 70%.
  • Architected applied LLM workflows targeting intelligent summarization, automated sentiment analysis, and systematic action-item generation.
  • Deployed distributed, highly-available ML pipelines using Celery and Redis to effortlessly manage fluctuating peak workloads.

Tech Stack

Core & Frameworks

Python FastAPI Flask LangChain

AI & Machine Learning

Transformers Whisper FAISS NLP Deepgram

Data & Infrastructure

PostgreSQL Redis Docker Celery GCP AWS