Computer Science undergraduate at VNR VJIET (CGPA: 9.1) focused on engineering highly scalable internal business applications, RESTful web services, and automated data pipelines using Java, Python, and cloud infrastructure.
$ status --current
> 3rd Year B.Tech CSE @ VNR VJIET (CGPA: 9.1 / 10.0)
$ current --workplace
> Software Developer Intern @ MyClassBoard (FastAPI & ONNX Computer Vision)
$ target --role
> Summer 2027 SWE, Backend Systems & Applied AI Internships
$ stack --primary-matrix
> Java (Spring Boot) | Python (FastAPI) | PostgreSQL | Docker | ONNX
Select an architectural topic below to activate the audio synthesizer.
VNR VJIET • B.Tech Computer Science (2024–2028)
Java 21 Virtual Threads & 4.7ms Caffeine latency
ONNX Runtime & 512-dim facial embeddings @ MyClassBoard
University Webathon (Post-Op Guardian)
Strict data isolation & unit/integration tests authored
Designathon 2026 (PhishGuard ML)
MyClassBoard (AgentP Platform)
FORTE (ForteHR Platform)
Co-Author: Someshwar Joshi • Proceedings of ICACECS 2025
Co-authored peer-reviewed research analyzing multi-variable sensor vectors for real-time water quality indexing. Evaluated the algorithmic convergence of Exponential Forgetting Factor Bit Error Rate (EF-BER) combined with spatial clustering and K-Nearest Neighbors (KNN) classification algorithms to filter environmental sensor telemetry.
High-Throughput Enterprise & Academic Platform
A high-throughput enterprise academic & internship tracking engine scaled to 785+ req/sec using Java 21 Virtual Threads (Loom), Caffeine in-memory caching, and optimized JPQL fetch joins.
Healthcare Recovery & Clinical Monitoring
Full-stack post-surgery recovery monitoring platform automating patient check-ins and clinical vital status reporting, securing 1st place overall at the university Webathon.
AI Threat Detection & Browser Security
Real-time browser security extension providing client-side malicious URL and download threat classification backed by Scikit-learn models and AI explanations.
From the initial catalyst to production computer vision pipelines at MyClassBoard, hackathon victories, and academic publications.
Developing asynchronous RESTful backend services in Python and FastAPI for AgentP. Engineered an ONNX Runtime + OpenCV computer vision pipeline (5-point facial landmark alignment, 512-dim embeddings) achieving 0% false-positive match rate. Scaled backend throughput via PostgreSQL pooling and authored 500+ Pytest tests.
Scaled backend concurrency using Java 21 Virtual Threads (Loom) and fine-tuned HikariCP pooling to sustain 785+ req/sec (0.00% error rate). Slashed listing latency by 99.1% (~531ms to ~4.7ms) via Caffeine in-memory caching (@Cacheable). Eliminated N+1 bottlenecks with JPQL fetch joins and DTO projections, with Flyway migrations and stateless JWT security.
Automated real-time phishing threat detection across URLs and file downloads via custom browser security extension (TypeScript, Manifest V3). Integrated hybrid ML algorithm pipeline (Scikit-learn on 50,000+ malicious records) with Gemini AI for human-readable threat explanations.
Engineered a full-stack recovery monitoring application to automate post-surgery patient check-ins, securing 1st place overall at the university Webathon. Built scalable REST APIs to process and validate critical patient recovery metrics for medical staff.
Co-authored peer-reviewed research paper on algorithmic water quality assessment utilizing EF-BER, clustering, and KNN algorithms. Evaluated classification accuracy across multi-parameter sensor vectors.
Architected ForteHR, an internal full-stack AI-powered HR analytics platform (React, TypeScript, Supabase) with RBAC. Designed an asynchronous LLM processing pipeline using the OpenAI API and custom system prompting, implementing semantic caching to reduce token usage and API overhead.
The turning point. Out of depth and failed completely—the necessary catalyst that inspired an obsession with mastering scalable architecture, rigorous database modeling, and production systems.
SWE Intern @ MyClassBoard • VNR VJIET
1d • 🌍
🚀 Excited to share an update on my work with AgentP at MyClassBoard! We engineered a computer vision pipeline with ONNX Runtime and OpenCV using 5-point facial landmark alignment and 512-dimensional embeddings—driving our false-positive match rate down to 0% across high-volume school event photos. Scaled via FastAPI and PostgreSQL pooling! #ComputerVision #FastAPI #Python #Engineering
Webathon Winner 1st Place • VNR VJIET
2w • 🌍
🏆 Thrilled to share that our team won 1st Place Overall at the university Webathon! We built Post-Op Guardian, a recovery monitoring system automating patient check-ins and clinical alerting for post-surgery teams. High-pressure building is always the greatest teacher! #winner #hackathon #healthcare #webdev
Backend & Distributed Systems
3w • 🌍
⚙️ Shipped Campus Project & Internship Tracker! Scaled backend concurrency using Java 21 Virtual Threads (Project Loom) and fine-tuned HikariCP pooling to sustain 785+ req/sec (47,000+ req/min) with 0.00% error rate. Slashed query latency from ~531ms to ~4.7ms (99.1% drop) via Caffeine in-memory caching (@Cacheable), eliminated N+1 bottlenecks via JPQL Fetch Joins & DTO projections, and enforced strict 3NF with Flyway! #Java21 #SpringBoot #ProjectLoom #PostgreSQL #Caffeine
Available for internships, technical discussions, and scalable backend collaborations. Let's build what scales.