Sagar Tripathy

Building from scratch for true understanding

Applied AI Engineer

Summary

Software Engineer with 6+ years of experience pivoting to AI/ML — drawn by its transformative potential as the defining technology of our era. Built LLMs from scratch, implemented LoRA fine-tuning, and developed AI agents through hands-on projects. Brings engineering rigor, client-facing experience, and team leadership from enterprise software delivery.

Technical Skills

ML/AI: PyTorch, Hugging Face Transformers, LLM Fine-Tuning (LoRA), Model Distillation, AI Agents, RAG
GPU/Systems: CUDA, Parallel Programming
Languages: Python, SQL
Tools: Git, Jupyter, Google Colab, REST APIs (Postman, ReadyAPI)

AI Projects

Nano-GPT from Scratch

  • Implemented GPT architecture from first principles in PyTorch
  • Extended to multi-modal variant handling text and images

LLM Fine-Tuning & Distillation

  • Implemented LoRA (Low-Rank Adaptation) for efficient fine-tuning
  • Performed model distillation on Qwen2.5-Coder
  • Domain adaptation fine-tuning on telecom data

AI Agents

  • Built AI agents from first principles
  • Developed Text2SQL agent for natural language database queries

PMPP - GPU Programming

All projects with code: tripathysagar.github.io/sagaTrip

Professional Experience

Software Test Engineer | Amdocs | July 2019 – Present

  • Led team of 2; handled onboarding, client calls, debugging, and integration
  • Built automation framework from scratch; managed 500+ test cases
  • Delivered solutions for telecom clients: LGI, Vodafone Italy, Maxis, CWC

Education

M.Tech, Computer Science & Engineering
NIT Calicut | 2017–2019 | GPA: 7.5
Thesis: Obstacle Detection, Depth Estimation and Warning System for Visually Impaired People
Published at IEEE TENCON 2019 | 14 citations

B.Tech, Information Technology
IIIT Bhubaneswar | 2012–2016 | GPA: 7.2

Learning

  • fast.ai — Practical Deep Learning for Coders