Hi, my name is
Preetham Dandu.
I'm a software engineer and researcher specializing in engineering the end-to-end lifecycle of production-grade AI systems. Currently, I'm a Master's student in Computer Science at Stony Brook University.
About Me
Hello! I'm Preetham, a passionate developer with a knack for creating elegant solutions to complex problems. My journey into tech began with a curiosity for how things work, leading me to a Bachelor's in Computer Science and Engineering from Vellore Institute of Technology, where I specialized in Networking and Security.
Today, I'm honing my skills as a Master's student at Stony Brook University. I've had the privilege of working as a Research Assistant, where I designed real-time ETL pipelines, and as a Software Engineer, diving deep into data science and machine learning.
Here are a few technologies I've been working with recently:
- Python
- Java & Spring Boot
- React.js
- Cloud Computing(AWS)
- Docker & Kubernetes
- LLMs (Llama-2 & RAG)
- PyTorch
- Distributed Systems
- SQL (PostgreSQL)
Work Experience
Research Assistant, Knowledge Systems Lab
Stony Brook UniversityJune 2024 - May 2025
Optimized ETL pipelines for large-scale datasets using Python & SQL.
- Engineered and deployed production-grade, real-time ETL pipelines to process and store live clinical fetal heart rate (FHR) data, establishing a mission-critical, HIPAA-compliant infrastructure enabling advanced AI research.
- Authored and validated a robust data anonymization framework to de-identify sensitive physiological signals, achieving 100% HIPAA compliance in collaboration with clinical partners and compliance officers.
- Developed a novel multi-scale LSTM model with an attention mechanism to predict fetal acidosis (pH ≤ 7.15), achieving 83.6% sensitivity and 0.82 AUROC on benchmark datasets.
- Pioneered an interpretable AI framework that links attention weights from deep learning models to Large Language Models (LLMs), generating automated, guideline-based clinical interpretations via structured prompt engineering.
- Optimized data workflows and established secure storage protocols, supporting reproducible research and multi-institutional data sharing.
Software Engineer
HCLTechAug 2023 - Dec 2023
Executed end-to-end data science projects, from data cleaning to predictive modeling.
- Led end-to-end data science lifecycles, including data cleaning, feature engineering, and predictive modeling.
- Developed and deployed machine learning models using R and TensorFlow to solve business problems.
- Generated actionable insights and visualizations to support data-driven decision-making for stakeholders.
Full Stack Developer Intern
Oriana ITJan 2023 - May 2023
Developed the e-commerce web app 'FashionBOLTz' with secure payments.
- Built the 'FashionBOLTz' e-commerce platform from scratch using React.js and modern web standards.
- Integrated secure payment gateways and implemented user authentication systems.
- Optimized frontend performance and component rendering, boosting navigation speed by 60%.
Software Engineering Intern
Matchday AIJan 2022 - Oct 2022
Collaborated with Star Sports and ISL clubs to build performance analysis tools.
- Developed computer vision pipelines using OpenCV to track player movements in real-time.
- Collaborated with Star Sports and ISL clubs to design custom performance analysis tools.
- Built backend services with Flask and SQLite to store and retrieve match analytics data.
Projects & Case Studies
RadianceAI Forecasting Engine
Problem: Solar energy output is volatile. How can we accurately forecast solar irradiance to optimize PV system efficiency?
- Python
- Llama-2
- RAG
- PyTorch
- FAISS
Photonimbus Photo Sharing App
Problem: Modern photo-sharing apps require scalable data storage, secure authentication, and a highly responsive user interface.
- Django REST
- Angular 8
- Google Cloud
- Docker
COVID Tracking Dashboard
Problem: Visualizing the real-time spread of a pandemic requires processing vast datasets with minimal latency for public use.
- Spring Boot
- Angular 9
- PostGIS
- GCP
Citation Prediction Count
Problem: Accurately forecasting the future impact of scientific papers is a complex challenge for researchers and institutions.
- Python
- Machine Learning
- Neural Networks
Plant Disease Detection
Problem: Farmers in remote areas lack quick, accessible tools to diagnose plant diseases, leading to potential crop loss.
- Django
- Deep Learning
- Distributed Systems
RECALL - Secure Manager
Problem: Standard note-taking apps often lack robust security and efficient cross-device access for sensitive information.
- PHP
- MySQL
- AWS
- Security
TuneGenie — AI-Powered Music Recommender
Problem: Help users instantly generate personalized playlists that match their mood, activity, and language—reducing the time and effort needed to discover music that "fits the moment."
- Python
- Streamlit
- Spotify API
- scikit-surprise
- OpenAI
- LangChain
SmartNotes - Cross-Device Universal iOS App
Problem: Modern users need a secure, high-performance note-taking app that seamlessly syncs across iPhone and iPad devices while maintaining data privacy and delivering smooth 60fps performance even with large note collections.
- Swift
- UIKit
- Core Data
- Combine
- Biometric Auth
- AES-256
Skills & Technologies
AI & Machine Learning
- Python (PyTorch, TensorFlow)
- LLMs (Llama, OpenAI) & RAG
- Computer Vision (OpenCV, YOLO)
- Scikit-learn, Pandas, NumPy
Backend & Systems
- Java (Spring Boot)
- Node.js & Express
- PostgreSQL, MongoDB, Redis
- Docker & Kubernetes
Cloud & Tools
- AWS (EC2, Lambda, S3)
- Google Cloud Platform
- Git & CI/CD Pipelines
- Linux & Bash Scripting