Portfolio

Kanishk Bhati

Full-Stack Engineer

I build production-grade systems and web applications, with a focus on backend architecture, distributed services, reliability, and optimization.

View Projects Download Resume

Or read my engineering notes on session validation, retries, and scheduling.

  • Spring Boot
  • PostgreSQL
  • Redis
  • Distributed Services
  • Timefold
Career

Experience

  1. Software Development Engineer (Full Stack)

    Aug 2024 – Present

    28One · Gurugram, Haryana

    Backend and full-stack work on scheduling and optimization services, REST API design, containerized deployments, and database query performance.

  2. Software Development Engineer Intern (Full Stack)

    Feb 2024 – Jul 2024

    Coding Ninjas · Gurugram, Haryana

    Full-stack feature work, front-end performance improvements, and automation of leaderboard updates using cron jobs and event-driven triggers.

Work

Projects

Intelligent Auto Planner

Constraint-Based Scheduling Engine

Professional work at 28One · Aug 2024 – present · Backend engineering

Scheduling engine integrated into an inspection platform to automate inspector-to-booking allocation. The planner evolved from a heuristic allocation strategy into a constraint-based optimization approach using Timefold.

Why this problem is hard

Assigning inspectors to bookings is a combinatorial problem rather than a sorting problem. Each assignment changes what the remaining ones cost, because travel time between sites, inspection duration, and per-inspector schedule and capacity limits all interact. A nearest-resource rule picks a defensible inspector for each booking in isolation and still produces a poor overall plan. Approved bookings also change, so the planner has to recompute without treating existing commitments as disposable.

Planning flow

Planning evolution

  1. Version 1 — Heuristic / Greedy Allocation

    • Nearest-resource assignment strategy
    • Location-based prioritization using geospatial filtering
    • Re-planning of previously approved bookings

  2. Version 2 — Constraint-Based Optimization

    • Integrated Timefold for vehicle-routing optimization
    • Solver-driven allocation across scheduling constraints
    • Modeled capacity limits and unassigned-booking penalties, and configured solver termination strategies (time-based and unimproved-spent limits)
    • Exposed REST endpoints for dynamic planning and re-planning
    • Improved the planning approach by modeling constraints explicitly rather than relying only on heuristic assignment

Real-world constraints

  • Travel time
  • Inspection duration
  • Schedule limits
  • Capacity limits

Outcome

Schema and query optimization on this work reduced data-retrieval latency by 35%.

Tech

  • Spring Boot
  • PostgreSQL
  • Docker
  • GitHub Actions
  • Timefold

Built at 28One. Source code and internal system details are not publicly available.

MarketMate

Distributed Marketplace Platform

Personal project · Full-stack · Live in production

Classifieds platform (OLX-style) for buying and selling used goods. Built as a distributed, multi-service system with separate frontend, backend, admin, and real-time chat services, deployed across custom domains with SSL and CDN-backed frontends.

Why this is technically interesting

The marketplace features were not the hard part. Two independently deployed services — the Spring Boot API and the NestJS chat service — both have to trust the same user session. That meant validating session state fast enough for a frequently executed path, without routing every authentication decision in the chat service through the main API, and without treating a cache miss as proof that someone is logged out.

Architecture

Core application
  • api.marketmatecloud.in — Spring Boot REST API
  • app.marketmatecloud.in — Angular SSR frontend
  • admin.marketmatecloud.in — Nx-based admin
  • Spring Boot API and Angular SSR frontend share an Nx monorepo
Services
  • Chat service — NestJS real-time messaging
Delivery & infrastructure
  • Custom subdomains for api / app / admin
  • DNS and SSL configuration
  • CORS and CSP configured for production
  • CDN-backed frontend delivery
  • Backend and SSR on Render

Engineering Highlights

Authentication & session management

JWT authentication via httpOnly cookies, with PostgreSQL session persistence and Redis-based session validation and caching.

Real-time communication

Separate NestJS chat service for real-time messaging.

Resilience

Resilience4j retry and circuit breaker patterns for backend integrations.

Media & storage

AWS S3 for media storage.

Transactional email

AWS SES (SDK v2) for OTP, email verification, and user notifications, with domain-level DKIM setup.

Containerization

Dockerized services with multi-stage builds and Docker Compose orchestration.

Production security & delivery

Custom domains, SSL, CORS, CSP, and CDN-backed frontend delivery.

Observability

Prometheus and Grafana monitoring.

Outcome

Running in production across four deployed services and three custom subdomains, with Prometheus and Grafana dashboards and Resilience4j retry and circuit breaker policies on backend integrations.

Tech

  • Angular
  • Spring Boot
  • NestJS
  • Nx
  • PostgreSQL
  • Redis
  • AWS S3
  • AWS SES
  • Docker
  • Docker Compose
  • JWT
  • Render
  • Netlify
  • Prometheus
  • Grafana
Live Demo Repos: Frontend Backend Infra Chat

Alphabet Music API

Early project · REST API with Android client

Music management platform consisting of a Node.js/Express API and Android client, with MongoDB used for persistent music data. The API provides authentication and backend services consumed by the mobile application.

Engineering Highlights

Authentication & API access

JWT-based authentication for protected API access.

Music data management

Designed API flows for uploading and managing music metadata and persisted data.

Mobile integration

Connected the Android client with backend APIs and the MongoDB persistence layer.

File handling

Implemented multipart file handling with Multer for music uploads.

Tech

  • Node.js
  • Express.js
  • MongoDB
  • Multer
  • JWT
  • Java
GitHub: API Android App
Writing

Engineering Notes

Write-ups on backend problems I've actually run into — debugging, authentication, reliability, and the architectural decisions and trade-offs behind each fix.

Authentication & Backend Architecture

JWT + Redis Session Validation

A valid JWT doesn't mean a session should still be trusted. Adding Redis-backed session validation closed that gap without giving up the speed of stateless tokens.

Read Note

Reliability Engineering

Retry vs Circuit Breaker vs Fallback

Retries, circuit breakers, and fallbacks answer different questions when a dependency starts failing — picking the wrong one can turn a blip into an outage.

Read Note

AWS & Backend Architecture

S3 Presigned Uploads

How direct-to-S3 uploads reduce backend load, with retry handling and backend fallback for failed uploads.

Read Note Coming soon

Optimization & Backend Systems

Constraint-Based Scheduling

How a scheduling problem evolved from heuristic allocation toward constraint-based optimization using Timefold.

Read Note Coming soon

Performance Engineering

Database Query Optimization

Lessons from investigating slow APIs, query efficiency, indexing, and reducing database-related latency.

Read Note Coming soon

About

About Me

I'm a full-stack engineer focused on building production systems that solve real engineering problems.

My work spans backend architecture, API design, distributed services, authentication, scheduling and optimization, and production infrastructure. I enjoy working on problems where reliability, performance, and maintainability matter as much as getting the feature working.

The work I find most interesting is usually diagnostic. On MarketMate, session validation was functionally correct but querying PostgreSQL on nearly every request; the fix was deciding what belonged on the common path and what belonged in the fallback. On the Auto Planner, a greedy assignment strategy produced schedules that looked reasonable per booking and poor overall, which is what pushed the problem toward a constraint solver.

I'm particularly interested in understanding why a system behaves the way it does, finding the bottleneck or architectural constraint, and then designing a solution that is practical to operate and maintain.

Skills

Tech Stack

Backend

  • Spring Boot
  • NestJS
  • Node.js
  • Express.js
  • REST APIs
  • JPA
  • Hibernate
  • JWT
  • Java

Database & Cache

  • PostgreSQL
  • MongoDB
  • Redis

Optimization & Scheduling

  • Greedy Algorithms
  • Vehicle Routing Problem (VRP)
  • Timefold

Infrastructure & DevOps

  • Docker
  • Docker Compose
  • Multi-stage builds
  • GitHub Actions
  • Maven
  • AWS S3
  • AWS SES
  • Resilience4j

Deployment & Observability

  • Render
  • Netlify
  • Custom domains & SSL
  • CORS/CSP
  • Prometheus
  • Grafana

Frontend

  • Angular
  • TypeScript
  • SCSS
  • Nx
Contact

Open to Backend and Full-Stack Engineering Roles

I’m looking for work on backend-heavy systems — APIs, distributed services, scheduling and optimization, authentication, and performance. I’m also happy to talk through a specific technical problem without a role attached.

If you’d like to get in touch, feel free to reach out through email or connect with me on GitHub and LinkedIn.

Email GitHub LinkedIn