krish_lakhani
Resume

Ahmedabad, India

Krish Lakhani

Full-Stack Developer & AI Engineer

I build full-stack products and the AI systems that power them, from RAG pipelines to cloud-integrated backends, shipped end to end.

View work→Start a conversation
About/ 01

I build software that holds up in production, not just in a demo.

My work spans full-stack engineering and applied AI: building RAG pipelines that reason over real content, integrating monitoring and ticketing platforms like Splunk, Datadog, and ServiceNow, and shipping secure, cloud-deployed applications end to end.

I care about the parts that don't show up in a screenshot: encryption and access control on stored data, query performance under load, and system design that holds up once real users show up. That's shaped by hands-on experience across monitoring integrations, secure full-stack apps, and real-time systems.

Focus
Full-Stack · RAG · Cloud
Discipline
Computer Engineering (B.Tech)
Availability
Open to work
Services/ 02

A few ways I can help.

01

Full-Stack Web Applications

End-to-end web apps: React or Next.js frontends, Node.js or Django backends, and a database like PostgreSQL or MongoDB, built and deployed as one working product.

02

AI & RAG Integration

Retrieval-augmented chatbots and search: embeddings, vector databases, and LLM-powered features that answer from your actual content instead of generic responses.

03

Secure Backend & APIs

Authentication, encryption, and role-based access done right: JWT auth, OTP verification, and REST APIs built to hold up under real traffic.

04

Cloud Deployment & DevOps

Getting your app live and keeping it up: AWS or GCP deployment, Docker containers, CI/CD pipelines, and monitoring integrations with tools like Splunk and Datadog.

Selected work/ 03

Four systems, shipped end to end.

episteme
Gemini · RAG
reading page
01

Episteme

RAG-powered Chrome extension for verifying article claims

Problem

Readers have no fast way to tell whether an article's claims are original, already contradicted elsewhere, or just repeating what's already been said, without leaving the page to go check.

My role

Built the full RAG pipeline inside a Chrome extension: LLM-based thesis extraction, 768-dimensional embeddings, and pgvector cosine search against a Supabase vector store, plus an in-page AI chat for follow-up questions.

Outcome

An extension that classifies the 3 nearest related articles as supporting, contradicting, or extending the current thesis, verifying every claim against verbatim source text before it's persisted.

Stack
  • Chrome Extension (MV3)
  • Node.js
  • Vercel
  • Supabase
  • pgvector
  • Gemini
  • RAG
vaultify
idle
drop file to upload
02

Vaultify

Encrypted file vault with role-based access control

Problem

Teams needed a place to store and share files where security wasn't an afterthought: encryption, verified identity, and access limited to the right role, not just a login screen.

My role

Engineered the storage schema and security layer end to end: S3-backed uploads, AES encryption at rest, JWT authentication, OTP email verification, CAPTCHA, and role-based access control.

Outcome

A containerized vault that has handled 500+ uploads through an encrypted, access-controlled pipeline, deployed with automated CI/CD via GitHub Actions.

Stack
  • Node.js
  • Express.js
  • MongoDB
  • AWS S3
  • AWS Lambda
  • GitHub Actions
  • JWT
polistore
AR · Unity
align tag in frame
03

PoliStore

AR-based hardware inventory system built for Madhya Pradesh Police

Problem

Police units were tracking equipment by hand, with no fast way to locate, verify, or audit hardware across facilities.

My role

Built the full-stack inventory portal (secure JWT login, PostgreSQL-backed records) and the AR scanning experience in Unity for QR and AR-tag item tracking.

Outcome

A scalable inventory system where officers scan a tag for an instant, auto-updating item record. The project reached the SIH 2024 finals, ranking top 3 of over 1,000 teams.

Stack
  • MERN Stack
  • PostgreSQL
  • Unity
  • Augmented Reality
  • JWT
chatsphere
Socket.IO
connecting
04

ChatSphere

Real-time chat app built to hold up under concurrent load

Problem

A chat product is only as good as its worst moment: many people talking at once, without dropped messages or stale sessions.

My role

Built the real-time messaging layer with Socket.IO and WebSocket handling, JWT authentication, and MongoDB Atlas for session management.

Outcome

A chat app supporting 100+ simultaneous users with real-time delivery and fast session and query performance backed by MongoDB Atlas.

Stack
  • Node.js
  • Express.js
  • MongoDB Atlas
  • Socket.IO
  • WebSocket
  • JWT
Capabilities/ 04

What I reach for, by discipline.

A

Languages

The languages I build in day to day.

  • Java
  • C/C++
  • Python
  • SQL
  • JavaScript
  • HTML/CSS
  • PHP
B

Frameworks & Libraries

For building full-stack products end to end.

  • Node.js
  • Express.js
  • React.js
  • Django
  • MongoDB
  • Supabase
C

Cloud & Tools

Shipping, monitoring, and keeping things running.

  • AWS
  • GCP
  • Docker
  • Git
  • Postman
  • Linux
  • ServiceNow
  • Splunk
  • Datadog
  • Dynatrace
  • pgvector
D

Concepts

The fundamentals everything else sits on.

  • OOP
  • DSA
  • DBMS
  • OS
  • REST APIs
  • Computer Networks
  • LLM
  • RAG
  • Agile Methodologies
Contact/ 05

Have a system worth building? Let's talk.

Or reach me directly

Email me