Bardawal Bihari
Hello, I'm

Bardawal Bihari

Final-year M.Tech (CSE – AI & ML) student with strong hands-on experience in Large Language Models (LLMs), NLP, and Deep Learning. Skilled in fine-tuning transformer models (LoRA/QLoRA), building AI pipelines, and optimizing inference for real-world applications. Currently working as a Data Science Intern at PowerSchool India.

About Me

I specialize in applied AI with a strong mathematical understanding of machine learning and deep learning models. My interests lie in LLM fine-tuning, NLP pipelines, generative AI applications, and scalable AI systems. I enjoy building end-to-end solutions that move beyond research into production.

  • Location: Telangana, India
  • Email: bardawal5555@gmail.com
  • Phone: +91 9014803908
  • Status: Open to AI / ML roles

Technical Skills

Expertise & Tools

Programming

Python (Advanced), SQL, C/C++, Java (Basics)

AI & ML

Machine Learning, Deep Learning, Neural Networks, Model Mathematics

NLP & LLMs

Transformers, BERT, Instruction Tuning, LoRA / QLoRA

Generative AI

Prompt Engineering, RAG (Foundations), LLM Applications

Frameworks

PyTorch, TensorFlow, Scikit-learn, Hugging Face

Data Engineering

Pandas, NumPy, ETL Pipelines, Feature Engineering

Backend

REST APIs, Flask, FastAPI (Basics)

Systems

Linux Basics, Docker (Basics), Git, GitHub

Education & Experience

M.Tech – CSE (AI & ML)

VIT Vellore | CGPA: 9.07 | 2024 – 2026

B.Tech – CSE (AI & ML)

JNTU Hyderabad | CGPA: 7.97 | 2020 – 2024

Data Science Intern – PowerSchool India

Jul 2024 – Present

Built data pipelines, performed EDA and statistical analysis, supported ML model training, fine-tuning, and optimization using Python, Pandas, NumPy, and SQL.

Projects

LLM-Based Domain-Specific AI & Chatbot

Fine-tuned transformer-based LLMs using instruction tuning and PEFT (LoRA/QLoRA). Built NLP pipelines for NER, text classification, and semantic search, achieving ~85% accuracy. Developed low-latency LLM applications using Gradio.

LLMs NLP Transformers

AI-Driven Computer Vision Access Control

Implemented a deep learning-based face recognition system using OpenCV with cloud-based inference. Deployed using AWS Lambda focusing on real-time performance, scalability, and reliability.

Computer Vision Deep Learning AWS

Get In Touch

Let's connect and create something amazing