🧠
📊
🦀
🐍
VK
👋 Hi, I'm Vijay Karanjkar

Software, Machine Learning,
& Artificial Intelligence

M.S. Computer Science @ ASU (4.0 GPA) — Building high-performance systems, ML pipelines, and full-stack applications.

About Me

Who I Am

VK

M.S. Computer Science

Arizona State University

GPA: 4.00

Open to opportunities

I'm a Software Engineer pursuing my M.S. in Computer Science at Arizona State University with a perfect 4.0 GPA. My work spans across distributed systems, ML pipelines, and full-stack engineering.

From architecting inference servers at BARC handling 100K+ requests, to building event-driven microservices at Tikaj, I bring a systems-thinking approach to every problem. I thrive at the intersection of performance-critical backends and intelligent data systems.

When I'm not engineering systems, I'm exploring quantitative finance, competitive programming on LeetCode, or contributing to open-source projects.

4.0
GPA
4+
Internships
3+
Projects
🧠

AI & ML

PyTorchScikit-LearnXGBoostLightGBMGNNsTensorRTRAGLLMs
💻

Languages

PythonRustGoJavaTypeScriptC++SQL
☁️

Cloud & DevOps

AWSDockerKubernetesTerraformCI/CDKafkaRedis
🌐

Web & Data

ReactNext.jsNode.jsFastAPISparkPostgreSQLAirflow
Experience

Where I've Worked

Bhabha Atomic Research Center (BARC)

Research Project Intern

Mumbai, IndiaFeb 2024 – May 2024
  • Architected ML inference pipelines via Triton Inference Server and Docker to handle 100k+ requests with 99.9% uptime
  • Optimized neural models with TensorRT, boosting inference speed by 40% and cutting latency to ~0.2s on 1.2M images
  • Built an NLP data pipeline using LLMs and Fairseq, generating 50k+ synthetic samples and improving BLEU score
TritonTensorRTDockerLLMsPyTorch

Tikaj

Full Stack SDE Intern

RemoteSept 2023 – Dec 2023
  • Built event-driven systems using RabbitMQ and MongoDB, scaling sequential data processing for 1,000+ sources
  • Integrated Redis caching to reduce database reads by 60%, achieving sub-0.2s latency for high-traffic APIs
  • Implemented real-time notification pipelines, delivering instant alerts across 1,000+ data watch sources
RabbitMQMongoDBRedisREST APIsNode.js

IIIT Vadodara

Research Intern

Gujarat, IndiaMay 2023 – July 2023
  • Built a RAG-based code security analyzer using LLMs to detect OWASP Top 10 vulnerabilities across 5+ repos
  • Designed a document ingestion pipeline parsing 10,000+ lines of code into vector embeddings
  • Achieved 82% precision on injection and authentication-related OWASP issues
RAGLLMsOWASPVector DBPython

Extent

Frontend Developer Intern

RemoteDec 2022 – Jan 2023
  • Developed 6 pixel-perfect React and TypeScript dashboards, cutting bug resolution time by 10%
  • Built responsive SASS web applications ensuring cross-device compatibility across 3 breakpoints
  • Refactored frontend architecture to improve reusability, reducing duplication by 25% across 6 dashboards
ReactTypeScriptSASSUI/UXREST APIs
Projects

What I've Built

🦀
Rust

TCP Message Queue

A high-performance Rust TCP broker built with Tokio, sustaining 10+ Gbps throughput and 100+ concurrent connections with zero-copy fan-out.

  • 10+ Gbps throughput, 100+ concurrent connections
  • 87% memory reduction via zero-copy fan-out
  • 40% faster parsing with 32-byte binary protocol
RustTokioDashMapDocker
View on GitHub
🔍
Python

Credit Card Fraud Detection

End-to-end ML pipeline on IEEE-CIS dataset (1.1M rows). GNN on transaction-identity edges achieved 84% recall and $68.1M projected savings.

  • 97.50% accuracy with LightGBM on 1.1M rows
  • GNN boosted recall to 84%, FDR down to 12%
  • $68.1M projected savings from fraud detection
PythonPyTorchXGBoostGNN
View on GitHub
🔗
Go

URL Shortener

High-performance Go/Fiber REST API with Redis caching and rate limiting. Sub-50ms resolution for 5,000+ daily users with real-time analytics.

  • Sub-50ms link resolution for 5,000+ daily users
  • Configurable TTL-based link expiration
  • Real-time click analytics for 10,000+ links
GoFiberRedisDocker
View on GitHub