Arin Balyan's avatar
Flag of India
text-3xl text-zinc-950 font-medium

Arin Balyan

Arin Balyan — AI Engineer & Full-Stack Developer

About

AI Engineer building AI-powered products with modern web & ML stacks.

Skilled in Python, PyTorch, TensorFlow, FastAPI, React, and Vue.js for full-stack AI development.

Deployed scalable systems at 99.8% uptime with complete DevOps ownership (Docker, CI/CD, cloud infra).

Created @JobHunter — AI job tracking platform with resume parsing, interview scheduling, and real-time status tracking.

Built @Scrappy — flexible web scraping framework for structured data extraction from multiple sources with scheduling.

Engineered RAG systems generating 500+ study pages and 100+ quizzes at 92% accuracy, cutting content creation time by 85%.

GitHub Contributions

Stack90

Experience

M

Mediajenie

Current Employer

• Design and build production RAG pipelines — document ingestion, embedding model selection and benchmarking, pgvector storage, and hybrid SQL + vector retrieval over permission-scoped data via FastAPI.

• Harden retrieval reliability with chunking and retrieval-window strategies, hallucination guardrails, confidence scoring, semantic caching, and async LLM query handling through Redis + Celery.

• Own AI infrastructure end-to-end — LoRA/QLoRA fine-tuning on domain data, vLLM deployment with GPU tuning and failover, versioned prompt architecture, and Prometheus + Grafana observability across GPU, API, DB, and sync layers.

  • Python
  • FastAPI
  • RAG
  • pgvector
  • PostgreSQL
  • Redis
  • Celery
  • vLLM
  • LoRA/QLoRA
  • Docker
  • Prometheus
  • Grafana
B

BrilliED

  • FastAPI
  • Python
  • Vue.js
  • Supabase
  • Render
  • LangChain
  • PostgreSQL
  • Docker
  • CI/CD
  • Team Leadership
  • Performance Optimization
  • DevOps
S

SkillBit Technologies

  • Python
  • FastAPI
  • Machine Learning
  • Natural Language Processing
  • Sentiment Analysis
  • Pandas
  • Scikit-learn
  • RESTful APIs
  • Data Preprocessing
  • Team Collaboration
E

Education

  • Python
  • C++
  • SQL
  • JavaScript
  • HTML
  • CSS
  • Machine Learning
  • Deep Learning
  • Data Structures
  • Algorithms
  • Computer Vision
  • Natural Language Processing
  • Mathematics
  • Physics
  • Chemistry
  • Computer Science
  • English
  • Science
  • Mathematics
  • Social Science
  • English
  • Hindi

Projects(8)

Bulk job-board scraper with 100+ sites, email enrichment, deterministic quality scoring, and multi-format exports (CSV/JSONL/XLSX/Parquet). Designed for scheduled bulk-first operations with per-site rate limiting, proxy pools, and resume support. Scales from laptop to server.

  • TypeScript
  • Python
  • Web Scraping
  • Email Enrichment
  • CSV/JSONL/Parquet

Achievements(9)

IIM Nagpur's Arthneeti 4.0

Top 10

Secured Top 10 rank among 1,500+ participants in market analysis case competition.

2024-01

Zelestra x AWS Hackathon

Top 25 Global

Achieved Top 25 global ranking among 500+ teams building an ML solution in 48 hours.

2024-01

GSSoC Contributor

50+ Merged PRs

Contributed 50+ merged PRs across 15+ open-source projects.

2024-01

Raycast Unsplash - Windows Support

Open Source

Added Windows support for wallpaper, clipboard, and save functionality to the official Raycast Unsplash extension.

2025