Hi, I'm AI Software Engineer
Umesh Ghaskata
I build production-ready AI systems that turn research ideas into practical products. My work focuses on Generative AI, LLMs, RAG, and agentic applications, and I enjoy taking projects from prototype to deployment with clean engineering and cloud-native tooling.
About Me
I'm Umesh Ghaskata, an AI Software Engineer and Computer Science graduate from the University at Buffalo, with a background from IIT Bombay. I focus on building intelligent systems that are practical, reliable, and ready for real users.
My work spans Generative AI, LLMs, multimodal RAG, agentic AI, and cloud-native software engineering. At the DRONES Lab, I worked on NASA AIST-21 research in hyperspectral image classification and UAV path planning, and helped reduce edge inference latency and improve accuracy through FPGA-based optimization.
Beyond research, I've built end-to-end systems including multimodal RAG platforms, LLM fine-tuning pipelines, GenAI-powered web applications, and cross-platform mobile apps. I enjoy working across Python, PyTorch, FastAPI, AWS, React Native, vector search, and MLOps workflows to bring ideas to life.
Portfolio: umeshghaskata.com
Tech Stack
AI / LLM Stack
Education
University at Buffalo, SUNY
Master of Science (Computer Science)
2024 – 2026 | Buffalo, NY | GPA: 3.86/4.0
Distributed Robotics and Networked Embedded Systems (DRONES) Lab
IIT Bombay
Bachelor of Technology (Aerospace Engineering)
2020 – 2024 | Mumbai, India
Experience
Software Engineer
Roswell Park Comprehensive Cancer Center · Jun 2025 – May 2026 · NY, United States
- Fine-tuned an open-source LLM (Gemma3-1B) on an AWS Cloud Linux instance using PyTorch, Transformers and Quantized LoRA. Architected a personalized AI coaching chatbot using a (RAG), grounding the model's responses.
- Hosted fine-tuned LLM using vLLM to Private VPC, desgined Low latency API to eliminate reliance on OpenAI API.
- Build a custom MCP server acting as a adapter, securely connecting AI assistant directly to users' private NoSQL databases.
- Owned the end-to-end integration of the AI backend with a cross-platform React Native frontend, enabling real-time, personalized conversational coaching for users.
- Managed source code and collaborative development using Git, implemented CI/CD pipelines with GitHub Actions.
- Implemented Redis semantic caching and rate limiting in a Node.js Express backend, reducing redundant API requests.
- Followed Agile practices through weekly meetings with user, advisor : taking direction and constructive criticism.
Research Assistant
DRONES Lab (NASA AIST-21 Project) · Aug 2024 – Mar 2026 · NY, United States
- Trained, evaluated, and deployed Deep Learning models (PCA+MLP, FPGA-based architectures) for hyperspectral image classification improving land-cover recognition accuracy from 75.4% to 94.5%.
- Reduced edge inference latency from 295ms to 52ms using optimized Model to enable real-time classification.
- Engineered benchmarking DL model across edge devices RTX 4060, Jetson Orin, and Xavier guided Hardware selection.
- Published Thesis on energy-efficient coverage path planning; developed a Python-based optimization framework.
Software Engineer
Scitara Corporation · Nov 2022 – Feb 2023 · Mumbai, India
- Developed low-latency RESTful APIs for SaaS Product using Docker and Kubernetes (AWS EKS), establishing a highly available infrastructure capable of supporting high-throughput AI inference workloads that reduced testing time.
- Facilitate backend with FastAPI Microservices with JWT Authentication and MockAPI service with PostgreSQL, equipped with Kubernetes(AWS EKS) for production.
Web Developer
IITB Rocket Team · Full-time · Mar 2022 – Jul 2022 · IIT Bombay
- Built and maintained the team website using HTML, CSS, and responsive layout techniques.
Projects
AI Shopping Agent
Personal Project |
Built an AI-powered shopping assistant that accepts natural language product queries and a budget, searches Google Shopping via SerpAPI, filters results by price, and uses an OpenAI model to compare the shortlisted products and recommend the best option with reasoning. The frontend is served by FastAPI and the backend is designed for public deployment on AWS EC2.
Tech Stack: Python, FastAPI, Uvicorn, OpenAI GPT-4o Mini, SerpAPI, Pydantic, HTTPX, Requests, python-dotenv, AWS EC2.
View Code Live Demo
Multimodal RAG Platform
Personal Project |
Built a production-ready Multimodal RAG platform that answers questions from PDFs, DOCX files, and images using semantic search and LLMs. Developed an end-to-end ingestion pipeline with OCR, image extraction, AI captioning, intelligent chunking, embedding generation, and vector indexing. Implemented context-aware retrieval using OpenAI Embeddings and Supabase pgvector with cosine similarity search, enabling accurate responses with document citations.
Tech Stack: Python, FastAPI, OpenAI GPT-4o, OpenAI Embeddings, RAG, Semantic Search, Vector Embeddings, Supabase, PostgreSQL, pgvector, OCR, PyMuPDF, python-docx, Pillow (PIL), Pydantic, Docker, HTML, CSS, JavaScript.
View Code Live Demo Video DemoDeepface Similarity Engine
IIT Bombay |
Built a facial recognition app using VGGFace (ResNet50) and MTCNN for face detection and embedding. Stored 2048-feature vectors using Pickle. Deployed a web app using Streamlit for real-time celebrity classification.
Tech Stack: Python, TensorFlow, Keras, VGGFace (ResNet50), MTCNN, OpenCV, NumPy, Scikit-learn, Streamlit, Pillow (PIL), Pickle.
View Code Live Demo
GenAI-Powered Career Readiness Platform
Personal Project |
Built a career preparation platform that uses Generative AI to analyze resumes against job descriptions, identify skill gaps, and generate personalized interview preparation plans. Leveraged Google Gemini to create ATS-friendly resume content, technical/behavioral interview questions, candidate-job match scores, and structured learning roadmaps.
Tech Stack: React.js, Node.js, Express.js, MongoDB, Google Gemini, JWT, Zod, Axios, Puppeteer, PDF-Parse.
View Code Live Demo
Supervised Fine-Tuning Gemma 3-1B LLM with LoRA vs QLoRA
Personal Project |
Experimented with parameter-efficient fine-tuning (PEFT) by training Gemma 3 1B IT on a custom Smoking Cessation / Motivational Coaching dataset (~1,500 conversations) on AWS EC2 (Tesla T4). Compared LoRA and QLoRA (4-bit NF4 quantization) trade-offs using BERTScore and ROUGE-L metrics. Both methods significantly outperformed the base model.
View Code View Model
Image-Caption-Generator
University at Buffalo |
End-to-end Deep Learning-based Image Caption Generator that automatically generates natural language descriptions for images. Combines Computer Vision and NLP using a pre-trained CNN for feature extraction and an LSTM-based sequence model for caption generation, learning to produce meaningful captions word-by-word from input images.
View Code
Autonomous Drone for Search Operations
IIT Bombay |
Developed an autonomous drone using Pixhawk + RPi, HOG-based detection, and A* algorithm with ultrasonic obstacle avoidance.
View CodeML Projects: Logistic Regression, CNNs & RL
SUNY Buffalo |
Implemented logistic regression from scratch using gradient descent and L2 regularization. Built and tuned CNNs and MLPs in PyTorch with dropout, batch normalization, LR scheduling. Designed a custom RL environment and trained agents using SARSA and Double Q-learning.
View Code
Coordinated UAVs for Efficient Spraying
IIT Bombay |
Modeled a 2D grid-based spraying path using OpenCV and HSV heatmaps. Applied Traveling Salesman Problem to optimize waypoint sequence and simulated A* and Dijkstra algorithms for flight efficiency.
View Report
Operating Systems Programming
IIT Bombay |
Built a Linux shell using fork(), exec(), wait() and implemented dynamic memory management in xv6. Used pthreads, mutexes, and semaphores for multi-threaded synchronization in user-space programs.
View Code
Get in touch.
Tell us what you're building and where you want to go next.
Prefer email?
umeshghaskata@gmail.com ▶Start a conversation.
Fields marked * are required.