I'm Suhas, a Software Developer.

Master of Science in Computer Science and Engineering

University at Buffalo, SUNY

GPA 3.67 / 4.0 · Machine Learning, Deep Learning, Data Models & Query Languages, Algorithms, Computer Security, AI for Health

Bachelor of Engineering in Information Science and Engineering

New Horizon College of Engineering

GPA 3.74 / 4.0 · Graduated with distinction

Software Development Intern

Jacobs School of Medicine, University at Buffalo

• –Developed a HIPAA-compliant health literacy app in React Native (Expo) with an Express backend, covering auth, data storage, and offline sync.
• –Delivered the full sign-in/sign-up flow, from app screens through account creation and session handling, using AWS Cognito with token rotation and revocation.
• –Engineered insert-only audit logging that captures 100% of PHI access events in a separate database, meeting HIPAA audit-trail requirements.
• –Implemented AWS KMS encryption for patient personal and health data, and built a CMS that lets researchers update app content directly, removing content updates from the engineering release cycle.

Research Aide

ESC Lab, University at Buffalo

• –Built a multi-page responsive website for the lab's BIRA summer research program using React.js and Tailwind CSS, with separate sections for each program's details and schedule.

Research Assistant (Volunteer)

University at Buffalo

• –Evaluated security compliance of LLM-generated code (LLaMA 3.1 and GPT) against 12 CWEs across 350+ code samples through manual code review.
• –Benchmarked each model's adherence to Secure Coding Practices and compared CWE violation patterns to identify recurring security weaknesses in LLM-generated code.

Associate Engineer

Brillio · CTO Team

• –Spent most of my time here building a multi-agent RAG pipeline that automatically converted legacy enterprise code (TIBCO BusinessWorks) into modern .NET microservices, achieving 70% code conversion accuracy which saved a lot of manual rewriting effort.
• –Built a Text-to-SQL chatbot over a 100K-row ServiceNow incident database using LangChain and Claude via AWS Bedrock so non-technical teams could ask questions in plain English instead of writing SQL.
• –Improved the internal developer portal so new engineers could spin up cloud deployments in 15 minutes instead of 40 minutes.

Project Intern

Hindustan Aeronautics Limited

• –Built the data management system for a UAV project in C++, handled communication between the main computer and the drone's sensors, reducing packet loss by 15%.
• –Designed a real-time GUI for sensor monitoring so engineers could track system status without manually reviewing logs.

A health literacy mobile app focused on women's reproductive health. Built with React Native (Expo) on the frontend and an Express + Prisma backend on AWS (Elastic Beanstalk, RDS PostgreSQL, S3, Cognito), with offline-first architecture, HIPAA-compliant audit logging, and column-level encryption for sensitive user data. I built the app's screens, backend services, authentication, and data security architecture.

An AI system that watches meal videos of parents and kids and detects food intake events. Finetuned R3D-18 Video ResNet on 135 videos (709K labeled frames) hitting 73.1% validation accuracy. Integrated YOLO to crop the child region before inference, which improved eating detection confidence 6x.

A fashion-entertainment iOS + web platform where users tag outfits to TV shows and movies. I handle the entire backend AWS infrastructure, batch content uploads, search with tag disambiguation, and shipped the iOS app via TestFlight.

Internal knowledge platform that uses GPT-2 to suggest solutions when someone logs an incident. Built the backend APIs and integrated the model reduced average resolution time by 25%.

Mobile app built for Alzheimer's patients that records conversations, identifies who's speaking, and narrates summaries back to the user. Used FaceNet for face recognition and Pyannote for speaker diarization got 85% speaker ID accuracy. This became my IEEE paper.

Real-time visual aid using YOLOv3 to detect and classify objects in live video feeds for visually impaired users — 92% detection accuracy.

UCSF QBI Hackathon 2026

UCSF · 2026

Built HiveBlot, a VLM-based pipeline that extracts structured data from western blot figures in scientific papers and enables natural-language search over the results via text-to-SQL.

Brillio Silver Star Award

Brillio · April 2025

Awarded for strong performance and delivering results despite hurdles — recognized specifically for commitment to Excellence and Innovation.

AWS Certified Cloud Practitioner

Amazon Web Services · June 2024

Certified on core AWS services, cloud architecture fundamentals, security, and pricing models.

IEEE ICONAT 2023 Publication

IEEE International Conference for Advancement in Technology (ICONAT)

Published research on an assistive technology app for Alzheimer's patients — facial recognition and speaker diarization to help them recall conversations.

Toycathon — National Level Hackathon

Government of India · 2022

Selected as one of the top 117 winners from over 1.2 lakh participants across India. Built a digital board game in Python around Indian culture and knowledge.