Open to opportunities

Sheshan S Maniyur

@sheshansmaniyur

Message

I build offline AI applications, agentic RAG systems, and production machine learning pipelines.

India

Message

What I'm looking for

I'm looking to apply my project-backed experience in AI engineering, agentic systems, RAG pipelines, and machine learning in a professional environment.

I've built AI systems including an offline YouTube video summariser at Vivarttana Technologies, eliminating cloud inference costs with a Faster-Whisper speech-to-text pipeline and BART-Large-CNN summarisation. I delivered a privacy-first Flask application that produces URL-to-summary results with on-device storage.

Through my M.Tech research and self-driven work, I've engineered climate-risk forecasting for 87 pan-India cities and an agentic RAG platform unifying 391 documents across eight source types. I work with Python, LangGraph, FastAPI, PostgreSQL, pgvector, ChromaDB, Docker, hybrid search, and LLM evaluation to build grounded, reliable AI products.

Experience

Work history, roles, and key accomplishments

Offline YouTube Video Summariser Intern

Vivarttana Technologies

Oct 2025 - Jan 2026 (3 months)

Built a fully offline speech-to-text pipeline for YouTube videos, reducing cloud costs to zero and achieving 95% transcription accuracy. Integrated BART-Large-CNN for summarization and shipped a privacy-first Flask web application.

Python FFMPEG BART Large CNN Flask Speech to Text Summarization Docker

Intern

Vivarttana Technologies

Oct 2025 - Jan 2026 (3 months)

Built a fully offline speech-to-text pipeline for YouTube video summarisation, reducing cloud costs to zero and achieving 95% accuracy. Integrated BART-Large-CNN for long-form transcript summarisation and shipped a privacy-first Flask web application.

Python FFMPEG BART Large CNN Flask Speech to Text Docker Summarization