Skip to content
View bhagyarana's full-sized avatar
😎
Exploring & Learning
😎
Exploring & Learning

Block or report bhagyarana

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
BhagyaRana/README.md
╔══════════════════════════════════════════════════════════════╗
β•‘                                                              β•‘
β•‘                    B H A G Y A   R A N A                     β•‘
β•‘        Software Engineer Β· System Designer Β· Builder         β•‘
β•‘                                                              β•‘
β•šβ•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•β•

Β  Β  Β  Β 



Typing SVG

$ whoami

I'm a Software Engineer at Wells Fargo who cares deeply about system design, developer experience, and shipping things that matter. My superpower is translating complex financial infrastructure into clean, scalable frontends β€” and then going home to build AI products from scratch.

I write about distributed systems and big tech architecture on my blog β€” topics like how Uber handles 40M req/s with caching, how Google Docs resolves concurrent edits, and why your database is not a message queue.

  • 🏦 Day job: Frontend architecture on the Policy Governance Platform β€” 1000+ production PRs, projected $11M+ annual savings
  • πŸ›  Building: FUMONEY β€” AI financial planning for Indian salaried professionals
  • ✍️ Writing: System design deep dives β€” Uber caching, Netflix streaming, distributed databases, and more
  • πŸŽ“ NIT Surat Β· B.Tech CS Β· CGPA 9.52 Β· Silver Medalist Β· Dept. Rank 2
  • πŸ† AMDOCS HackFest 2022 β€” All India Rank 2 out of 11,000+ participants
  • β™Ÿ Chess & Basketball β€” State-level athlete

$ cat stats.json

πŸš€ Production PRs πŸ’° Business Impact πŸŽ“ CGPA πŸ† Hackathon Rank
1000+ $11M+ 9.52 / 10 AIR 2


$ git log --projects

πŸ’° FUMONEY

AI-powered financial planning for India

Built for Indian salaried professionals who have no idea if they're on track for retirement. FUMONEY turns messy financial inputs into a personalized Financial Fitness Score, retirement corpus projections, and real-time what-if simulations.

Hard engineering problems solved:

  • Designed a goal-based planning engine that models compound growth, inflation-adjusted returns, and multiple simultaneous financial goals without state explosion
  • Built real-time simulation system with debounced recalculation to stay responsive under rapid user input
  • Auth + RLS via Supabase to ensure zero cross-user data leakage at the DB level

Next.js 14 TypeScript Supabase PostgreSQL Redis OpenAI API

Live

πŸ€– AI Cohort

Build production-ready AI agent systems

An open-source learning platform teaching developers to build stateful, scalable AI agents β€” not toy demos. Covers LangChain, RAG pipelines, LangGraph orchestration, and Model Context Protocol with real deployment patterns.

Hard engineering problems solved:

  • Architected a multi-agent orchestration layer using LangGraph where agents hand off tasks with typed state β€” no spaghetti tool-calling chains
  • Implemented hybrid retrieval (semantic + BM25 keyword) for RAG to handle both conceptual and exact-match queries without two separate pipelines
  • Designed MCP server integration patterns that work across local and cloud-hosted agent deployments

LangChain LangGraph RAG MCP Vector DBs React

Live Source

πŸš€ BuildSpace

Ship real products over a weekend

A marketplace where startups post bounty projects and builders collaborate in-person to ship working products in 2–3 days. Handles project listings, builder profiles, team formation, and bounty payouts.

Hard engineering problems solved:

  • Built optimistic UI updates with server-state sync so builders see changes instantly without stale cache issues during fast-paced weekend sprints
  • Designed a matching algorithm that pairs builder skill profiles to project requirements without requiring explicit skill tagging by either party

Next.js 14 TypeScript React Tailwind CSS PostgreSQL

Live Source

🏦 Policy Governance Platform

Enterprise banking @ Wells Fargo

Frontend architecture for the system managing 10,000+ banking policies across Wells Fargo's core workflows β€” used daily by compliance, legal, and operations teams.

Hard engineering problems solved:

  • Led migration of critical policy rendering pipeline from legacy AngularJS patterns to a component architecture that cut load time significantly and eliminated an entire class of rendering bugs
  • Built a shared state management layer that synchronizes policy lifecycle state across 6+ concurrent user roles without race conditions
  • Drove 1000+ production contributions β€” features, defect fixes, performance optimizations β€” with zero P0 incidents

AngularJS Spring Boot SQL CI/CD Agile

Impact PRs


$ cat blog/latest.md

I write in-depth system design breakdowns β€” the kind that actually explain how things work, not just that they work.

πŸ“ Why Your Database Is Not a Message Queue The antipattern that silently kills production systems at scale β€” and what to use instead. 5 min Β· Feb 2026

πŸ“ Uber's Caching Strategy: 40M Requests Per Second The integrated caching layer behind Uber's scale β€” and the patterns worth stealing. 7 min Β· Feb 2026

πŸ“ Monolith to Microservices: The Parts Nobody Warns You About Distributed transactions, data consistency nightmares, and why Uber went back. 7 min Β· Feb 2026

πŸ“ How Netflix Streams to 230M Users Without Buffering Shot-based encoding, multi-region failover, and why their architecture is brilliant. 7 min Β· Feb 2026

πŸ“ Google Docs Real-Time Collaboration: The Algorithm Operational Transformation β€” how 50 people edit one doc simultaneously without conflicts. 7 min Β· Feb 2026

πŸ“ Distributed Databases Demystified: Cassandra, DynamoDB, BigTable LSM trees, consistent hashing, gossip protocols β€” stop treating NoSQL as a black box. 8 min Β· Feb 2026

πŸ“ Back-of-Envelope Calculations: The Senior Engineer Superpower The skill that separates senior engineers β€” estimate system capacity in 5 minutes. 7 min Β· Feb 2026

πŸ“ Consistent Hashing: The Algorithm Powering Half the Internet From DynamoDB to Discord β€” how one algorithm decides where your data lives. 9 min Β· Jan 2026


$ ls tech-stack/

Languages

JavaScript TypeScript Python Java C++ SQL

Frontend

React Next.js Angular Tailwind CSS ShadCN

Backend & Databases

Node.js Express Spring Boot PostgreSQL MongoDB Redis Supabase

AI / Agents

LangChain OpenAI LangGraph

Cloud & DevOps

AWS GCP Docker Vercel GitHub Actions


$ cat achievements.log

[2022] AMDOCS HackFest          β†’ All India Rank 2   |  11,000+ participants
                                   Built [describe what you built here] in 24 hours
[2023] NIT Surat Graduation     β†’ Silver Medalist    |  CGPA 9.52  |  Dept. Rank 2
[2021] Google Kick Start        β†’ Round C: 888th globally  |  Round G: 654th globally
[2021] JPMorgan Code for Good   β†’ Finalist (Top 8%)
                                   Built [describe the project here] for [NGO/org name]
[2021] Upraised Embark Fellow   β†’ Top 2% of 4,55,534 candidates
[2019] Dhirubhai Ambani Scholar β†’ Merit-based scholarship, NIT Surat

$ cat currently_thinking.md

Things occupying my engineering brain right now:

  • Agent reliability β€” Most LLM agent failures aren't model failures, they're orchestration failures. I'm exploring how LangGraph's stateful checkpointing changes the failure recovery story for long-running agents.
  • Financial data modeling β€” Building FUMONEY taught me how hard it is to model Indian financial products (NPS, ELSS, HRA) in a way that's both actuarially correct and explainable to a 26-year-old.
  • Read path optimization β€” After writing about Cassandra's LSM trees, I've been thinking hard about read amplification tradeoffs in systems where reads vastly outnumber writes.

$ neofetch --competitive-programming

CodeChefΒ  CodeforcesΒ  HackerEarth


$ tail -f activity.log

GitHub Streak

Minimal. Scalable. Production-Ready.

Pinned Loading

  1. Jarvis_Voice_Assistant Jarvis_Voice_Assistant Public

    [Just a Rather Very Intelligent System] This repository consist of Voice Assistant of Iron Man [aka Tony Stark]. Let's be honest, it's not as intelligent as in the movie, but it can do a lot of coo…

    Python 7 2

  2. POP_FLIX POP_FLIX Public

    One Stop Destination to Book Tickets

    EJS 2

  3. SUDOKU-SOLVER SUDOKU-SOLVER Public

    This Web App Generates and Solves Sudoku

    Python

  4. UrbanSpice.github.io UrbanSpice.github.io Public

    HTML