Anyone can ship for Day 1.
I build for Day 2.
Bipasha Bhattacharjee — backend & AI engineer, 3+ years shipping distributed systems.
Software Engineer with 3+ years building production-grade systems — full-stack products, cloud-native backends, and AI engineering.
Nowrebuilding this portfolio from the ground up — new system, evidence over adjectives
The Career Timeline
From CS fundamentals to AI engineering — every role left behind a system that's still running.
M.Tech — Artificial Intelligence
Graduate Research
Focused on machine learning, deep learning, and applied AI systems.
- Research across NLP, representation learning, and model evaluation
- Bridged academic ML with production engineering practice
B.Tech — Computer Science
Undergraduate
Built the fundamentals — data structures, distributed systems, and software engineering.
- Strong algorithmic foundation (800+ LeetCode problems solved)
- Global Rank 11 — Arcesium She Codes
Tricon Infotech
Software Engineer
Shipped full-stack product and platform features — from the UI down to the APIs underneath.
- Built React + TypeScript front-ends backed by FastAPI services
- Designed REST/event APIs and improved system observability
- Containerized workloads and shipped on Kubernetes
Finbox
Backend / Systems Engineer
Built backend systems for financial data and real-time decisioning.
- Event-driven pipelines with Kafka & RabbitMQ for reliable processing
- Microservices for low-latency data ingestion and scoring
- Hardened services for correctness, idempotency, and scale
New Engen
AI / Backend Engineer
Shipped AI features and backend services for a marketing intelligence platform.
- Shipped LLM-backed workflows and agentic automations
- Cloud-native services across AWS & GCP with autoscaling
- Owned end-to-end delivery from design to production rollout
Systems I've Built
Production case studies: the problem, the architecture, and the trade-offs I'd make again.
File Processing Microservice
Async log processing with real-time analytics
Problem
Large log files are slow and risky to process inside a request — they block the client and can time out. The service needed to accept big uploads, process them in the background, and surface live analytics safely.
Key Decisions
Queue-based processing
BullMQ workers process uploads asynchronously so the API stays responsive under load.
Real-time dashboard
A Next.js 15 + React 18 UI streams job progress and extracted log statistics (errors, keywords).
Containerized stack
Docker Compose packages the API, worker, and Supabase-backed storage for reproducible runs.
Impact
AI Lab
Generation engines, agent workflows, evals, and a prompt library — built and tested like production code, not notebook demos.
Text Generation
TextGen — Neural Text Generator
A natural-language generation app that produces coherent text from custom prompts, combining Bi-LSTM/LSTM networks with a pre-trained LLM.
- Bi-LSTM + LSTM architecture trained and served with PyTorch
- Attention mechanisms + prompt engineering for contextual coherence
- Interactive Streamlit UI for live generation
Agent Workflows
AI-Powered Course Generator
A multi-agent FastAPI backend that turns a one-line course idea into a complete curriculum, broken into modules with real lesson content.
- Multi-agent architecture: research → modules → lesson content
- Production-grade async FastAPI backend
- Brief description in, full structured course out
Applied ML
ECG Arrhythmia Classifier
A full-stack app that loads ECG data, visualizes arrhythmia events, and classifies new signals with a trained ML model.
- Interactive plots with event markers (AFIB, VTACH, BRADY…)
- Prediction endpoint classifies new ECG signals
- Python ML backend with a web visualization layer
NLP
Spell Checker
A Python text-processing tool that detects and corrects spelling errors using algorithmic and linguistic analysis.
- Algorithmic detection + correction of spelling errors
- Linguistic analysis to improve readability
- Lightweight, dependency-free core
Skills Universe
Five orbits — Frontend, Backend, Cloud, AI, and DevOps. Select a domain to explore the stack within.
Frontend
Grouped by where it's actually been used, not a number I assigned myself.
Engineering practices
Developer Dashboard
Open-source footprint, pulled live from GitHub — nothing here is a repeated stat.
Public repositories
GitHub stars earned
GitHub followers
Achievement Wall
Numbers, not adjectives — rankings, problems solved, and systems shipped.
Global Rank 11 — Arcesium She Codes
Top 11 worldwide in a competitive algorithmic challenge.
800+ LeetCode Problems Solved
Sustained algorithmic practice across DS, DP, graphs, and more.
3+ Years in Production
Shipping resilient systems across fintech, AI, and platform teams.
M.Tech in Artificial Intelligence
Advanced study in ML, deep learning, and applied AI systems.
Sub-100ms Decisioning
Real-time streaming platform serving low-latency decisions.
20+ Systems Shipped
Full-stack products, microservices, and AI workflows in production.
Playground
Small interactive tools — because the best way to understand a system is to play with it.
Big-O Visualizer
n = 20
JSON Formatter
{
"name": "bipasha",
"stack": [
"python",
"react"
]
}Open Source
Tools and libraries I've shared with the community.
Let's Build Something
Have a system that needs designing, scaling, or rescuing? I'd love to hear about it.