Hi, I’m Mayank Mahajan.
Building production-ready web applications and scalable backend systems.
I design and build web applications with a strong focus on backend architecture, APIs, databases, authentication, real-time systems, and cloud deployment.
Who I Am
A personal look at how I think about software and engineering.
I'm a backend-focused full-stack developer who enjoys turning ideas into working software. My journey started with building web applications, but I quickly found myself drawn to what happens behind the interface — how APIs are structured, how databases scale, how authentication remains secure, and how background jobs process data reliably.
Today, my work centers on designing modular REST APIs, structuring relational and document databases, implementing caching and rate limiting, building real-time event systems with WebSockets, and integrating AI workflows with LLM tool calling.
I believe in building software that is clean, maintainable, and straightforward. I value strong engineering fundamentals — understanding time and space complexity, designing predictable database schemas, writing testable code, and containerizing services for reproducible deployments.
Outside of direct development, I am an Information Technology student at HVPM COET, where I actively coordinate campus tech fests, organize coding competitions, and lead technical workshops for fellow students.
Quick Background
What I Do
How I contribute across the software engineering lifecycle — from database design and APIs to full-stack applications and cloud deployment.
Backend Engineering
Designing modular REST APIs, authentication systems, database schemas, caching layers, background task workers, and real-time services.
Full-Stack Development
Building end-to-end web applications with Next.js, React, TypeScript, and modern frontend tooling, with seamless backend integration.
Cloud & DevOps
Containerizing applications with Docker, configuring automated CI/CD workflows, working with Linux environments, and deploying backend services.
AI-Powered Applications
Building applications that integrate LLMs, real-time token streaming, agent execution loops, structured tool calling, and workflow automation.
Projects I’ve Built
Real-time competition engines, campus ERP platforms, developer agent runtimes, and async data pipelines.
CTF Platform
Real-Time Competition Platform
A full-stack Capture The Flag competition platform built for running live cybersecurity tournaments with real-time solve broadcasts and dynamic score updates.
A modular Node.js and Express backend with Socket.IO for live event delivery, Redis Pub/Sub for state synchronization across clients, MongoDB persistence, and a responsive Next.js frontend.
Eliminating race conditions during concurrent challenge submissions by combining atomic database updates with dynamic scoring decay formulas and token-bucket rate limiting.
PUB/SUB channel "ctf:events:live"
⚡ FIRST BLOOD: user "hex_hunter" solved pwn-300
↳ Redis dynamic decay: 500pts → 482pts recalculated
Socket.IO broadcasted to active concurrent sockets
Token-bucket rate limiter: zero false positives
OPENCLAW
AI Coding Agent Runtime
A developer agent runtime that takes high-level coding goals, executes terminal commands, inspects logs, and modifies code with human-in-the-loop safety.
A stateful agent loop with dynamic plan generation, JSON schema tool calling via OpenRouter, a rich terminal UI, and a bidirectional Telegram bot bridge for remote status monitoring.
Implementing reliable human-in-the-loop permission gates for sensitive shell commands and file edits, preventing destructive actions while maintaining an autonomous reasoning loop.
> Plan: 3 steps generated for goal "diagnose crash"
Step 1: Read error logs [SUCCESS]
Step 2: Inspect memory leaks [GATED: USER APPROVED]
Step 3: Apply patch & run regression tests [PASS]
↳ Telegram Bridge: notification delivered to mobile
TransactIQ
AI-Powered Transaction Processing
An asynchronous transaction processing pipeline designed to ingest, validate, and reconcile large CSV financial batches with AI-assisted anomaly detection.
A FastAPI microservice coupled with Celery distributed workers, a Redis message broker, PostgreSQL database with Alembic migrations, and Gemini AI inference for transaction classification.
Decoupling heavy file ingestion and reconciliation from the HTTP request cycle into background worker queues with automated health checks to prevent server blocking.
$ celery -A app.worker worker --loglevel=info
[tasks] ingest_csv_batch, anomaly_eval, reconcile
✓ Task ingest_csv_batch[9f8e2] processed asynchronously
⚡ Gemini AI inference: 0 anomalies flagged
↳ DB Pool: 8 connections active • Redis Broker: 0 pending
Neuron AI
Multi-Model AI Application
A multi-model AI conversational workspace supporting persistent chat sessions, conversational branching, and real-time streaming token responses.
A Next.js App Router application consuming real-time token streams via OpenRouter, with Better Auth authentication and Prisma ORM backed by PostgreSQL.
Architecting an incremental streaming interface that provides immediate visual feedback while managing branching conversation histories and session security.
> Model: Multi-model switching via OpenRouter
Streaming responses deliver tokens incrementally
Better Auth session verified (JWT + Prisma)
Branch history: multi-fork conversation thread
Interactive streaming UI with zero layout shift
How I Progressed as an Engineer
A short story of how I developed my focus from frontend fundamentals to backend architecture and distributed systems.
Foundations & Core CS Rigor
Deep focus on core computer science foundations, Object-Oriented Programming (C++, Java), Linux environments, and solving data structure problems daily.
Full-Stack Development & Distributed Concepts
Transitioned into building production web platforms. Mastered Node.js/Express backends, Next.js full-stack architectures, relational schemas in PostgreSQL, document stores in MongoDB, and Redis caching.
Scaling Systems, Real-Time & AI
Engineered more complex applications: a real-time CTF competition engine with Socket.IO, asynchronous task queues with Celery, an AI coding agent runtime, and resilient full-stack architectures.
Continuous Engineering & Growth
Focused on building production-ready web applications, scalable backend systems, cloud deployment fundamentals, and continuous problem-solving on LeetCode.
Tech I Work With
A clear view of my primary full-stack development stack alongside supporting cloud, systems, and AI tooling.
Core Daily Technologies
Cloud & DevOps
Containerization, automated pipelines, deployment targets, and Linux environments.
Additional Technologies
Real-time WebSockets, algorithmic languages, mobile frameworks, and LLM APIs.
What I’m Working On
A live snapshot of the systems I’m building, the problems I’m solving, and the technologies I’m exploring right now.
Backend Services & APIs
Writing modular, type-safe REST APIs and WebSocket services using TypeScript, Express, and PostgreSQL.
Campus Tech Leadership
Organizing coding competitions, technical fests, and hands-on engineering workshops for students at HVPM COET.
Data Structures & Algorithms
Practicing algorithmic problem solving on LeetCode with continuous focus on Medium-tier DP, graphs, and trees.
Cloud & Infrastructure
Experimenting with Docker containerization, multi-service Compose configurations, and automated CI/CD pipelines.
AI-Powered Workflows
Exploring structured tool calling, streaming token pipelines, and agent loops using OpenRouter and LLM APIs.
Engineering Practice & Evidence
Tangible evidence of continuous coding: open-source repositories on GitHub and daily problem-solving on LeetCode.
Active personal and team repositories covering backend APIs, real-time WebSocket engines, and AI applications.
Consistent practice in space-time complexity optimization, tree/graph traversals, dynamic programming, and arrays.
Let’s Build Something
Interested in software engineering opportunities, freelance projects, or technical collaborations? Let’s connect.
Based in India (IST, UTC+5:30) • Open to full-time roles & freelance