Tejas Pawar

Master's Student in Data Science | AI Builder | Research Assistant

"Exploring Generative AI, Agentic AI, and building real-world AI products."

About Me

About Tejas Pawar

Director of SuNaAI Lab | Graduate Student in Data Science, University of Delaware

I'm Tejas Pawar — an AI builder and researcher passionate about Generative AI, NLP, and agentic systems. I created SuNaAI Lab as a space where ideas move fast, experiments turn into projects, and knowledge is shared openly.

My work explores how LLMs, semantic search, and memory-augmented systems can be transformed into real-world applications. From natural language understanding and dialogue systems to cloud-native AI deployment, I focus on creating tools that are not only technically sound but also impactful for learners, researchers, and industries.

At SuNaAI Lab, I wear many hats:

Engineer

Building and deploying AI projects with modern stacks (FastAPI, Docker, AWS/GCP).

Software Developer

Architect and sole builder of the SuNaAI Lab website and platform.

Researcher

Experimenting with NLP pipelines, retrieval systems, and multi-agent workflows.

Educator

Sharing insights through blogs, posts, and collaborative learning.

🌟

For me, SuNaAI Lab is more than a project — it's a vision for an open, experimental AI space where passion, curiosity, and innovation come together.

Experience

Graduate Assistant

2025–Present

University of Delaware – IT Academic Technology Services

As a Graduate Assistant, I contribute to the UD StudyAiDE project (ats.udel.edu/udstudyaide), an AI-powered platform designed to enhance teaching and learning at scale.

My responsibilities include:

•Developing agentic AI systems that enable autonomous and intelligent support for academic use cases.
•Fine-tuning large language models (LLMs) to specialize them for education-focused tasks such as summarization, topic review, and personalized assistance.
•Leveraging AWS Bedrock services to deploy, scale, and manage LLM-based applications in a secure cloud environment.
•Building and maintaining backend workflows that ensure smooth integration of AI capabilities into the StudyAiDE ecosystem.

This role allows me to apply Generative AI, NLP, and cloud-native deployment in a production-level academic platform, shaping how AI can be used responsibly in higher education.

Machine Learning Intern

Mar 2023 – Jul 2024

LogicMo Systems Pvt. Ltd.

During my internship at LogicMo Systems, I worked on a computer vision project where I focused on solving challenges related to video and image data processing.

Key contributions:

•Designed and trained computer vision models to analyze video streams and image datasets for object detection, classification, and feature extraction.
•Implemented end-to-end ML pipelines, including preprocessing (frame extraction, noise reduction, normalization) and feature engineering.
•Optimized model performance through fine-tuning, hyperparameter tuning, and evaluation using standard CV benchmarks.
•Applied PyTorch deep learning frameworks to deploy robust vision solutions.
•Documented workflows and experimental results, ensuring reproducibility and clarity for future engineering teams.

This role gave me strong practical exposure to computer vision, deep learning, and applied ML engineering, preparing me to tackle real-world AI challenges with unstructured multimedia data.

AI Hackathon Participant

Aug 2025

HensStreet Hacks 2025

Won Creativity & Originality Prize 🏆 for AR + GenAI project

Independent AI Researcher

2025–Present

SuNaAI Lab

Building M-Maze (memory-augmented assistant), DebateGPT (AI debate system), AR cybertext, and other AI research projects

Education

Master of Science in Data Science

2024-2026

University of Delaware

Newark, DE, USA

Bachelor of Engineering in Electronics & Telecommunication

2020-2024

Savitribai Phule Pune University

Pune, Maharashtra, India

Bachelor of Engineering Honors in Data Science

2022-2024

Savitribai Phule Pune University

Pune, Maharashtra, India

Projects

🧠

M-Maze

Memory-augmented AI assistant using Qdrant & agentic AI

Advanced AI system with long-term memory capabilities and intelligent agent workflows.

🎤

DebateGPT

Voice-based AI debate platform (AWS Bedrock + Polly + Transcribe)

Interactive debate system leveraging AWS services for natural language processing and speech synthesis.

🥽

AR Hackathon Project

Real-time 3D cybertext generator using LLM + Blender + A-Frame

Award-winning project combining generative AI with augmented reality for immersive text experiences.

🌐

SuNaAI Lab Platform

Collaborative AI lab website and project hub

Modern web platform built with Next.js, Tailwind CSS, and Docker for showcasing AI research and projects.

Skills

🤖
AI & Data Science

LLMs

RAG

Fine-tuning

Agentic AI

NLP

AR/VR AI

💻
Programming

Python

PHP

JavaScript

SQL

🛠️
Frameworks/Tools

FastAPI

Flask

Docker

AWS (Bedrock, S3, Lambda)

GCP (Vertex AI)

Achievements

🏆

Creativity & Originality Prize