Open to Backend / SWE Roles — Full-time or Internship

Hi, I'm
Jay Soni

Backend engineer shipping production systems. I've built real-time WebSocket servers load-tested to 90+ connections, cut API latency by 40–60% with Redis, and deployed 5 systems on AWS that are live right now. My GitHub work has attracted outreach from industry professionals — I'm ready to bring that same standard to a team.

5 AWS Deployments
90+ WS Connections
170+ DSA Solved
Download Resume ↓ Email Me → GitHub ↗ LinkedIn ↗
Jay Soni
FastAPI Django AWS PostgreSQL MongoDB Redis RabbitMQ Docker Celery WebSockets GitHub Actions SQLAlchemy

About

I build backends that work under real conditions

Most students build CRUD apps and call it done. I went further — real-time systems, message queues, Redis caching, role-based auth, CI/CD pipelines — and deployed all of it on AWS EC2 where it actually runs. Five projects. All live.

I care about writing backend code that doesn't fall apart when real traffic hits it. That means understanding why a query is slow. It means knowing when async helps and when it just adds complexity. It means building things that are easy for the next engineer to pick up.

Currently interning at KEP Labs (Bhopal) working on production APIs with Django, FastAPI, and PostgreSQL. Looking for my next role where I can take on more ownership and work on harder problems. Open to relocating anywhere in India.

🎯 Targeting Backend Engineer & SDE roles — Full-time or Internship
🏗️

Production

5 systems deployed on AWS EC2 with Docker, Nginx, and GitHub Actions CI/CD — live at my domains.

⚡

Tradeoff Thinking

Redis over schema denormalization. Sync over async when the complexity isn't worth it. Cookie sessions over JWTs when revocation matters. Tools chosen for Purpose.

☁️

Full Stack of the Backend

From API design to async queues (RabbitMQ, Celery) to database optimization to cloud deployment — I've touched every layer.

Tech Stack

What I work with

Languages

Python3 SQL

Frameworks & APIs

FastAPI Django Django REST Framework Flask

Databases & Caching

PostgreSQL MongoDB SQLite Redis

ORM & Data Access

SQLAlchemy Django ORM

Async & Messaging

Celery RabbitMQ WebSockets Async/Await

DevOps & Cloud

Docker AWS (EC2, S3) GitHub Actions Nginx CI/CD Linux

Templating & Frontend

Jinja2 HTML CSS

Experience

Where I've worked

Feb 2026 – Present Keshava Elite Projects
(KEP Labs) Active

Python Backend Intern

  • Built and maintained REST API endpoints using Django and FastAPI for internal tooling and business workflows
  • Worked with SQL databases for data modeling and query writing handling data flow and storage across API layers
  • Debugged API issues by tracing request flows and validating responses ensuring correct behavior and stability across endpoints
  • Tested APIs using Postman and pytest, validating outputs and data flow to confirm expected responses before delivery

Projects

Selected work — all deployed, all real

FinVerse — Finance Platform

↗
Deployed · AWS EC2

Full-stack financial management platform with role-based access control 3 roles, 17+ API endpoints, a Jinja2 dashboard, and a REST API layer, all sharing the same auth backend.

⚙ RBAC enforced at the FastAPI dependency layer even with the UI — access control holds even if someone hits the API directly
  • 3-role system (admin / analyst / viewer) with server-side enforcement on every protected route
  • Dual interface: browser dashboard via Jinja2 + REST API via /api/* — one shared auth layer for both
  • 17+ endpoints covering full CRUD, data import/export, and analytics views
  • Pytest suite validating RBAC logic and endpoint behavior — not just happy paths
17 API Endpoints RBAC at API Layer Dual Interface Live on AWS EC2
FastAPI SQLAlchemy PostgreSQL Jinja2 Pytest AWS EC2

Crypto Trading Simulator

↗
Deployed · AWS EC2

Real-time trading simulator with WebSocket-based price streaming, FIFO portfolio tracking, and live P&L load-tested to 90+ concurrent connections and deployed on AWS with full CI/CD.

⚙ Per-connection async handling with Redis as shared state avoids bottlenecks without blocking the event loop
📬 This project attracted direct outreach from a VIP & Partnerships Manager at Crypto.com Exchange — they cited the GitHub work as standing out
  • FastAPI WebSocket backend load-tested at 90+ concurrent connections no dropped connections, consistent message delivery
  • Redis caching for shared price + portfolio state reduced redundant DB reads and cut API latency by 40–60%
  • FIFO-based portfolio engine with full transaction history and real-time P&L accurate across hundreds of simulated trades per session
  • GitHub Actions CI/CD → Docker build → AWS EC2 deploy — push to main, it ships
90+ Concurrent WebSocket Connections 40–60% Latency Reduction Real-time P&L Engine
FastAPI Redis WebSockets Docker GitHub Actions AWS EC2

Secure E-Voting System 2.0

↗
Deployed · AWS EC2

Role-based voting platform built around correctness guarantees one vote per user, real-time result aggregation, and a layered architecture that keeps read performance fast under concurrent load.

⚙ Vote counting runs entirely in MongoDB aggregation pipelines server-side grouping with compound indexes, not pulling records into Python to count
  • Layered architecture: FastAPI → Services → MongoDB — deliberate separation so business logic, auth, and data access never bleed into each other
  • Compound indexes on voter ID + election ID fields — high-cardinality reads stay fast without full collection scans under concurrent load
  • Chose cookie-based sessions over JWTs here to keep revocation simple tradeoff made deliberately, not by default
  • bcrypt auth with duplicate-vote prevention enforced at the database level, not just application logic constraint holds even under race conditions
Compound Indexed for Concurrent Reads Vote Integrity at DB Layer Server-Side Aggregation
FastAPI MongoDB Cookie Auth bcrypt WebSockets GitHub Actions

Event-Driven Notification System

↗
GitHub

Async notification backend using RabbitMQ to decouple email delivery from the main request flow failed messages retry automatically, nothing gets silently dropped.

⚙ Manual acknowledgments + dead-letter queues — a message isn't marked done until delivery is confirmed, and failures go to a retry queue automatically
  • RabbitMQ producer-consumer workflow with durable queues and retry logic zero message loss on delivery failure
  • SendGrid integration for transactional email with delivery status tracking
  • PostgreSQL for user data, session state, and a full audit log of every notification dispatched
  • OAuth2 authentication with middleware-based request validation across all protected routes
Zero Message Loss OAuth2 Secured Full Audit Log
FastAPI RabbitMQ Celery SendGrid PostgreSQL OAuth2

Credit Approval Backend System

↗
GitHub

Loan approval backend with rule-based credit scoring, EMI calculation, and async processing via Celery built to mirror how a real fintech backend handles credit decisions.

⚙ ORM queries use select_related and prefetch_related throughout no N+1 queries, no unnecessary round-trips to the database
  • Django REST Framework APIs with optimized ORM queries deliberately avoided N+1 patterns from the start
  • Rule-based credit scoring engine: income, debt-to-income ratio, repayment history → eligibility category → interest rate
  • Celery + Redis for async loan computation heavy processing doesn't block the API response
  • EMI and amortization calculations per credit tier, stored and retrievable per user
  • PostgreSQL with transactional integrity financial data either commits fully or not at all
Async Loan Processing Rule-Based Scoring Engine N+1 Free ORM Transactional PostgreSQL
Django REST Celery Redis PostgreSQL Python

Achievements

By the numbers

5
Live AWS Deployments

Not demos — actual systems running on AWS EC2 behind Nginx, deployed via Docker and GitHub Actions CI/CD. You can visit them right now.

10+
AWS Educate Badges

Certifications across compute, databases, storage, networking, and security — not collected for the badge, used to make better infrastructure decisions on real projects.

170+
LeetCode Problems

DSA across arrays, trees, graphs, and dynamic programming — the kind of foundation that shows up when you're debugging a slow query or designing an efficient data structure under pressure.

Contact

Let's work
together

Open to Backend Engineer & SDE roles — full-time or internship. Willing to relocate anywhere in India or work remotely. I respond within 24 hours.

jaayysoni@gmail.com →
⌥
GitHub
github.com/jaayysoni
↗
in
LinkedIn
linkedin.com/in/jaayysoni
↗
☁
Credly
10+ AWS Educate Badges
↗
LC
LeetCode
170+ Problems Solved
↗
✆
Phone
+91 94078 20271
→