Swachhith

AI Engineer

Systems that observe,
decide and act.

Swachhith builds production agentic systems, retrieval pipelines and multimodal applications — the kind that run in front of real users rather than in a notebook. Based in Hyderabad.

Things I have built.

About this guy.

Outside the terminal: following Formula One with more optimism than McLaren has historically deserved, playing football badly and often, building small computer vision projects just to see whether they work, and taking photographs I never get around to sorting.

Where the work happened.

01

Valyouu Digital

Associate AI EngineerJul 2025 — PresentHyderabad

  • Built and deployed an AI-powered internal platform from scratch, integrating seven tools across research, content and marketing.
  • Architected agentic workflows with tool calling, memory and feedback loops in LangChain and LangGraph, orchestrating multi-agent systems at scale.
  • Designed multimodal RAG voice agents for marketing, deployed across five clients.
  • Automated video and design pipelines, fine-tuning LoRA models for generation across 10+ campaigns.
7tools integrated
30people using it
60hsaved per week

LangChain · LangGraph · RAG · LoRA

02

Kona AI

AI EngineerJan — Jun 2025Remote

  • Built an agentic system on LangGraph and the Playwright MCP server that explores a live application and writes its own test automation.
  • Generated Playwright test cases in TypeScript, widening coverage and cutting manual effort.
  • Created synthetic data to simulate diverse user scenarios, hardening product validation.
50%less test dev time
0specs written by hand

LangGraph · Playwright MCP · TypeScript

03

La Plateforme

Data Science InternApr — Sep 2024Marseille

  • Improved a chatbot for a leading French rail operator using NER, intent recognition and function calling over LLMs.
  • Leveraged LangChain and vector databases to sharpen semantic search and response accuracy.
  • Pre-processed multilingual datasets, enabling translation and lifting model performance.
  • Deployed and managed models on Microsoft Azure for scale and reliability.
95%response accuracy
4languages handled

LangChain · Vector DBs · Azure · NLP

04

Ernst & Young

Data Science InternJun — Aug 2023Hyderabad

  • Extracted actionable insight from large datasets with advanced programming and data visualisation.
  • Ran exploratory data analysis to surface patterns and outliers that informed strategic decisions.
  • Worked across functions to turn analysis into recommendations leadership acted on.
20%operational efficiency

Python · EDA · Visualisation