Siddhant Thete profile photo 1 on Tal

Siddhant Thete

Building with LLMs & RAG | LoRA Fine-Tuning · NLP · Python | CSE @ IIIT Guwahati | Open to Internships

Amravati

Siddhant’s AI Stack

codex
codex
cursor
cursor
opencode
opencode

Education

Indian Institute of Information Technology Guwahati

Bachelor of Technology - BTech, Computer Science

Indian Institute of Information Technology Guwahati

20234 yrs2027

Class 12, PCM

Maharshi Public School - India

20211 yr2022

Join Siddhant and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Siddhant Thete on Tal

Headline

Building with LLMs & RAG | LoRA Fine-Tuning · NLP · Python | CSE @ IIIT Guwahati | Open to Internships

About

I don't just study AI — I research it and ship it. I'm Siddhant, a third-year CSE student at IIIT Guwahati, working at the intersection of NLP research and applied Generative AI. ━━━━━━━━━━━━━━━━━━ WHAT I'M RESEARCHING ━━━━━━━━━━━━━━━━━━ I'm currently conducting research on machine translation under computational constraints, supervised by Dr. Rohit Tripathi. Key results so far: • BLEU 25.23 / chrF 50.87 on Hindi → Malay via LoRA fine-tuning of NLLB-200 Distilled600M • Recovered ~90% of full-dataset baseline using only 50% of the data and half the compute (83h vs. 132h) • Showed that quality filtering (length-ratio + character 3-gram Jaccard) can match 30% raw data performance with 40% less training time ━━━━━━━━━━━━━━━━━━ WHAT I'VE BUILT ━━━━━━━━━━━━━━━━━━ AskYourDocs — a live document intelligence system on HuggingFace Spaces. • Sub-2s latency • Zero infra cost • Handles PDF, DOCX, and TXT natively • Built on LangChain, FAISS, Groq LLM, and Streamlit • Deployed and serving real queries Argus — a container deployment and rollback automation system. • Detects failed deployments via health checks • Auto-rolls back to the last stable container • Built with Docker, Node.js, and MongoDB ━━━━━━━━━━━━━━━━━━ MY STACK ━━━━━━━━━━━━━━━━━━ • LLMs & GenAI: LangChain, RAG, LoRA fine-tuning, Prompt Engineering • NLP: Machine Translation, NLLB-200, BLEU/chrF/COMET evaluation • Languages: Python, Java, SQL • Infra: HuggingFace (Spaces, Transformers, PEFT), FAISS, Streamlit, Docker ━━━━━━━━━━━━━━━━━━ I'm in my 3rd year (graduating 2027) and actively looking for AI/ML/NLP internships for Summer 2026. If you're working on something in NLP, Generative AI, or intelligent systems — I'd love to connect. 📎 github.com/siddhant855 📩 siddhant.v.thete855@gmail.com

Location

Amravati

AI stack

codex

cursor

opencode

Education

Bachelor of Technology - BTech, Computer Science, Indian Institute of Information Technology Guwahati

Class 12, PCM, Maharshi Public School - India

Public links

github: https://github.com/siddhant855

Siddhant Thete - Building with LLMs & RAG | LoRA Fine-Tuning · NLP · Python | CSE @ IIIT Guwahati | Open to Internships in Amravati | Tal