SOFTWARE ENGINEER | AVID LEARNER | ARCHITECT ENTHUSIAST

BUILD.
CREATE.
SOLVE.

I build scalable backend systems, AI-powered applications and digital experiences.

HOMEKolkata, India
CURRENT LOCATION Pune | Bangalore, India
Rupam portrait
BACKEND
AI / ML
FRONTEND
DATABASE
SCROLL TO EXPLORE
01
ABOUT ME

MORE THAN
JUST CODE.

I'm Rupam Sadhukhan, a software engineer focused on building scalable systems, intelligent products, and practical digital experiences.

My work spans backend engineering, full-stack product development, AI-powered solutions, and enterprise-grade application design.

01

BUILD

Scalable systems and reliable software.

02

LEARN

Constantly exploring new technologies.

03

CREATE

Products that solve real problems.

04

DESIGN

Thoughtful and user-centric architecting of systems and interfaces.

JOURNEY

FROM SCHOOL TO BUILDING

2019 • 2021School

The Assembly of God Church School, Kolkata

ICSE 2019 — 90%
ISC 2021 — 93%
EntranceWBJEE Rank around 2K

Secured a strong engineering path and moved toward a technical career.

2022 • 2026College

Jalpaiguri Government Engineering College

CGPA: 9.00/10.00
InternshipPrismforce

Built hands-on software engineering experience in real-world product work.

CurrentASDE at Prismforce

Working on scalable systems, AI products, and production-grade engineering.

02
EXPERIENCE

WHERE I'VE
BEEN BUILDING.

Aug 2025 – Sep 2025
AMAZON MACHINE LEARNING SUMMER SCHOOL

TRAINEE (TOP 2.5%)

Bengaluru, India

  • Accomplished advanced training in ML, NLP, Deep Learning, and Computer Vision; applied statistical modeling and data-driven techniques to solve fraud detection and forecasting challenges, improving predictive accuracy by 18%.
  • Implemented and evaluated 10+ supervised and unsupervised models (classification, regression, clustering) using Python, Spark, TensorFlow/PyTorch, achieving 18% higher accuracy than baseline.
  • Implemented scalable ML pipelines for fraud detection and demand forecasting, improving prediction reliability by 92% and reducing errors by 18%.
Feb 2026 – Present
PRISMFORCE

ASSOCIATE SOFTWARE DEVELOPMENT ENGINEER

Pune, India

  • Engineered scalable backend services using Prisma ORM, SQL, and optimized database queries, improving reliability and performance for real-time analytics workflows.
  • Designed and integrated RESTful APIs with an Angular analytics dashboard, enabling scalable reporting workflows and seamless frontend-backend communication.
  • Developed a modular backend workflow integrating LangGraph and RAG services through API-driven architecture, improving response accuracy while reducing operational cost and maintaining scalable enterprise workflows.
BackendNode.jsSQLPrismaAIREST APIsAngular
03
SELECTED WORK

THINGS
I'VE BUILT.

01

Namma Food Platform

A scalable food delivery platform built with Java, Spring Boot, React.js, and AWS, serving 1K+ concurrent users with resilient backend services.

Java Spring Boot React.js AWS
↗
02

AIVA Mock Interviewer

An AI-powered interview simulation platform using NLP, Python, and Scikit-Learn to support adaptive interview workflows and reusable service components.

Python NLP Scikit-Learn AI
↗
03

Project Kshitij

A beach safety application built with Java, Kafka, Redis, and REST APIs to enable real-time alerts and predictive safety workflows.

Java Kafka Redis REST API
↗
04

Face Recognition - ComputerVision

A compact, easy-to-follow repository with utilities and notebooks to build and experiment with classical and deep learning face recognition pipelines.

Python OpenCV TensorFlow Computer Vision
↗
04
TECHNOLOGY

WHAT I
WORK WITH.

Java
Python
C++
JavaScript
SQL
HTML5
CSS3
Spring Boot
Java
Python
C++
JavaScript
SQL
HTML5
CSS3
Spring Boot
React.js
Node.js
Express.js
Django
REST APIs
JWT
Bootstrap
MySQL
React.js
Node.js
Express.js
Django
REST APIs
JWT
Bootstrap
MySQL
MongoDB
Redis
Prisma ORM
AWS EC2
Docker
GitHub
Postman
LangGraph
MongoDB
Redis
Prisma ORM
AWS EC2
Docker
GitHub
Postman
LangGraph
LangChain
RAG
LLaMA
Pandas
NumPy
Scikit-learn
Data Structures
Algorithms
LangChain
RAG
LLaMA
Pandas
NumPy
Scikit-learn
Data Structures
Algorithms
OOP
DBMS
Operating Systems
Computer Networks
Distributed Systems
System Design
OOP
DBMS
Operating Systems
Computer Networks
Distributed Systems
System Design
05
HOW I USUALLY INTERACT WHILE WORKING

THINK IN
SYSTEMS.

A notification service is never just a send button. I start with the user promise, make the failure modes explicit, and design the smallest system that can keep that promise in production.

01 / CONSTRAINTS

What must be true?

Urgency, ordering, consent, cost, provider limits and data sensitivity shape the design before code does.

02 / CONTRACTS

Where is the boundary?

Clear event schemas, idempotency keys and ownership rules keep teams from coupling through assumptions.

03 / OPERATIONS

How do we recover?

Every alert needs an owner, every retry needs a limit, and every migration needs a reversible path.

notification-service design session · online
LIVE REVIEW
We need email, SMS and push notifications. How would you make it reliable?
First, I would write down the user promise: accepted is not the same as delivered. Then I would map failure boundaries and let the API accept intent quickly while workers handle delivery.
What happens when a provider times out or sends the same event twice?
Idempotent consumers, bounded retries and a dead-letter queue. The event contract carries an idempotency key, correlation ID and delivery deadline, so recovery is deliberate instead of mysterious.
delivery-flow / decision-log REVIEWED
API→QUEUE→WORKER→PROVIDER
SLO measurable deliveryDLQ owned recovery
And how would you know it is working in production?
Measure before guessing: structured logs, queue depth, provider latency and alertable SLOs. I would ship one provider behind a feature flag, replay realistic failures, and expand only when the runbook and rollback path are proven.
That is how I usually work...
06
HAVE AN IDEA?

LET'S
BUILD.

Have a project, opportunity or just want to talk about technology?

TOP ACHIEVEMENTS
  • Amazon Machine Learning Summer School 2025 — Selected among the Top 2.5% candidates nationwide.
  • Smart India Hackathon 2024 Finalist — Ranked among the top 2.4% from 490K+ participants.
  • Build With India 2025 Semi-finalist — Selected among top 5000 from 25K+ participants.

I practice code like Max drives — not to finish the lap, but to shave milliseconds off the next one.

ENGINEERING MINDSET
RUPAM SADHUKHAN © 2026 BUILT WITH ♥ USING ANGULAR