Hi, I'm Shantnu
I build scalable backend systems and distributed architectures, craft full-stack, AI-powered applications, and turn complex problems into reliable, production-ready software.
Open to Opportunities
Actively seeking full-time Software Engineering roles. Let's discuss how I can contribute to your team.
About Me
Currently pursuing my Master's in Computer Science at Virginia Tech with a perfect 4.0 GPA. For me, tech has always been about diving deep into complex challenges and building intelligent solutions that truly matter. I'm constantly motivated to learn, build, and refine, whether it's debugging production issues or streamlining developer workflows. I find energy in the details that make systems more robust and efficient.
Bringing nearly 3 years of professional experience spanning Backend, Full-stack, AI Engineering, and Distributed Systems, I enjoy working end-to-end, whether designing architectures or integrating intelligent features. I've contributed to product cycles at every stage: taking something from 0→1, scaling it 1→N, modernizing legacy flows, or rapidly iterating on new ideas.
Backend Engineering
Designing reliable, high-performance backend services
Distributed Systems
Building scalable, fault-tolerant systems at scale
AI Engineering & ML
Applied ML systems from research to production
Full Stack & DevOps
End-to-end product delivery and deployment
Work Experience
My professional journey in software engineering
Research Assistant - ML Systems & Data Engineering
Technologies
Highlights
- Built a scalable misuse detection system processing 17M+ app-store reviews to identify privacy and consent concerns.
- Operationalized an embedding + regression pipeline (USE) to generate review-level risk scores and app-level aggregation for prioritization.
- Built distributed data processing workflows on AWS (S3/EMR/PySpark) for large-scale review ingestion and feature extraction, enabling repeatable experimentation and faster iteration cycles.
- Implemented few-shot LLM narrator labeling and moral analysis to convert flagged reviews into explainable, product-facing signals.
Projects
Research and engineering projects I've built
Featured Projects
OpenGPU | Distributed GPU Marketplace & Execution Platform
Technologies
Highlights
- Built a production-grade GPU rental platform enabling consumers to run workloads on remote GPUs with end-to-end trust, isolation, and payment guarantees.
- Developed a hybrid C++/TypeScript Windows desktop agent with NVIDIA NVML integration for real-time GPU monitoring, heartbeats, and live resource telemetry, achieving >99.5% node availability visibility.
- Built the web platform using Next.js 15 with multi-database architecture (PostgreSQL, MongoDB) and Redis caching for provider status management.
- Implemented fault-tolerant job orchestration handling client disconnects and retries, improving successful job completion rate by >30%.
- Integrated Stripe for escrow-style payments, releasing artifacts only after confirmation to eliminate producer-consumer disputes.
AI Powered Kiosk Restaurant Platform | Full-Stack Application
Technologies
Highlights
- Architected a kiosk and mobile application using React and TypeScript, enabling touch and voice-based interactions.
- Implemented QR scan-based login from mobile apps to kiosks, enabling personalized and streamlined user sessions.
- Developed a Node.js middleware layer to support API routing, authentication, and speech to text integrations via OpenAI Whisper.
- Integrated OpenAI GPT-4 API for intuitive, voice-driven navigation.
Currently Building
OutreachCopilot | AI-Powered LinkedIn Outreach Engine
Building a local-first AI application that generates personalized LinkedIn messages and InMails using structured inputs like resumes, job descriptions, and recipient profiles.
ReceiptSplit | AI-Powered Expense Splitting
Building an AI-powered receipt parsing and expense splitting application that combines OCR extraction with a human-in-the-loop review and correction flow.
More Projects
Developed a robust data pipeline to curate "NewsSlant," a corpus of 34,790 English tweets from 48 U.S. news outlets and 4M associated replies, with a strict chronological cap of the first 20 replies per thread (345,182 labeled replies) to mitigate viral outlier bias and manage computational cost.
Skills & Technologies
Technologies I work with
Languages
Tools & Frameworks
Databases
AI & ML Systems
Expertise
Education & Achievements
Academic background and recognition
Academic Journey
Masters in Computer Science and Applications
Virginia Polytechnic Institute and State University
Bachelor of Technology in Computer Science
Indian Institute of Information Technology Guwahati
Awards & Recognition
GreenHack 2021, Prague
Secured 1st place in the CSOB Track among 50+ teams.
Unihack 2021, Czech Republic
Awarded the Special Jury Award for innovative solutions.
Health-Hacks Hackathon, California, USA
Achieved 2nd Runner-Up position.
Skoda COVID Mobility Race 2020, Czech Republic
Earned the Kraj Vysocina Prize, competing against 20+ teams.
21-Day Data Challenge, Lighthouse Labs, Canada
Completed an intensive data challenge program.
Get In Touch
Let's discuss opportunities and ideas