Open to Spring 2027 Internships

Prakash
Pappala

I'm a Master's student in Computer Science at UT Arlington, currently doing research on LLMs and multimodal AI. Outside of coursework I build things agent pipelines, RAG systems, computer vision tools — and try to ship them properly, not just leave them as notebooks.

LLMs NLP RAG LangGraph Multimodal AI MLOps FastAPI Docker
Prakash Pappala
9
AI/ML Projects
96.65%
XGBoost Accuracy
99.99%
Sentiment API Confidence
3.67
GPA at UTA
projects

What I've Built

Agentic AI
Agentic Research Assistant
4-agent autonomous pipeline using LangGraph — Planner, Researcher, Critic, Writer working sequentially to produce structured research reports.
↳ Multi-agent orchestration · Docker · CI/CD
LangGraphGroq Llama3DockerGitHub Actions
Live Demo GitHub
RAG
RAG Document QA System
Upload any document and ask questions. LangChain + FAISS for semantic retrieval and Groq Llama3 for context-aware answers.
↳ Reduced hallucination vs vanilla LLM
LangChainFAISSDockerCI/CD
Full Stack ● Live
TimeBook — Student Productivity App
Full-stack web app for students to track time, take notes, and manage tasks. Live timer, custom categories, cross-device session sync, notebook, to-do list, and analytics by week, date, and year. Proposed as a pilot to UT Arlington.
↳ FastAPI + PostgreSQL + React · Deployed on Render & Vercel
FastAPIPostgreSQLReactJWT Auth
Live Demo GitHub
NLP
AI Resume Screening System
Matches resumes to job descriptions using cosine similarity with sentence-transformers and AI feedback via Groq Llama3.
Sentence TransformersGroq Llama3Streamlit
Multimodal
Multimodal CLIP Image-Text
Zero-shot image classification and cross-modal semantic search using OpenAI CLIP. Upload an image, match it to text labels.
CLIPHuggingFacePyTorch
REST API
Sentiment Analysis REST API
Production-ready FastAPI service with single and batch prediction endpoints using fine-tuned DistilBERT.
↳ 99.99% confidence on sentiment classification
DistilBERTFastAPIPyTorch
ML
Predictive Modeling — XGBoost
End-to-end ML pipeline comparing XGBoost and Random Forest with real-time prediction and feature importance visualization.
↳ XGBoost 96.65% vs Random Forest 94.35%
XGBoostScikit-learnStreamlit
LLM
AI Code Review Assistant
Multi-language code reviewer using Groq Llama3. Detects bugs, readability issues, and performance problems.
Groq Llama3Prompt EngineeringStreamlit
NLP
Knowledge Graph Builder
Extracts named entities and relationships from unstructured text using SpaCy and visualizes as directed graphs via NetworkX.
SpaCyNetworkXStreamlit
NLP
Text Classification — DistilBERT
Zero-shot text classification into custom categories using DistilBERT. No training required — define labels and classify.
DistilBERTHuggingFaceStreamlit

Tech Stack

AI / ML
LLMsNLPRAGTransformersFine-tuningPrompt EngineeringComputer VisionMultimodal AI
Frameworks
HuggingFaceLangChainLangGraphPyTorchScikit-learnXGBoostOpenCVFAISS
Backend / APIs
FastAPINode.jsExpress.jsREST APIsJWT AuthSocket.io
MLOps & Tools
DockerGitHub ActionsGitStreamlitGroq APIPandasNumPy
Languages
PythonJavaScriptSQLC
Frontend
React.jsHTMLCSS
experience

Work History

Volunteer Research Assistant
University of Texas at Arlington
Jan 2026 – Present
Built LLM pipelines using fine-tuned BERT/DistilBERT. Implemented RAG workflows with FAISS and embedding-based retrieval. Applied multimodal data fusion for cross-modal LLM analysis.
Technical Trainer & Lab Programmer
Siddhartha Institute of Technology and Sciences
Oct 2024 – Dec 2025
Delivered Python, Java, C, and DSA labs. Mentored students in debugging, OOP, algorithms, and software development fundamentals.
Teacher
Chaitanya Bharathi High School
Dec 2023 – Sep 2024
Taught CS fundamentals, algorithms, and programming using Python and C across multiple grade levels.

Get In Touch

I'm looking for Spring 2027 internships in AI/ML, software engineering, or applied research. Happy to chat if you're working on something interesting.