Open to backend & AI engineering roles

Krrish KumarBackend & AI Engineer

I build the systems behind intelligent products.

Backend systems, distributed architectures, and real-time infrastructure, with production AI on top. Founder of Parivar AI, live on Google Play.

View selected workGitHub(opens in a new tab)Résumé

system.map

clientgatewayservicesredisqueuellm

Selected work

One AI product people install, and three backend systems with public source. Open “How it's built” on any of them for the architecture and the trade-offs.

Live productAI product · Founder

Parivar AI

AI companion for Indian families.

An AI family companion app. When you can't call home, talk to Maa, Papa, or Dadi in Hindi, by chat or by voice call.

  • Live on Google Play
  • AI conversations in Hindi & Hinglish
  • Voice calls, photo reactions
  • App and backend built by me

Key technologies: Node.js · LLM APIs · Text-to-speech · Flutter

Visit Parivar AI(opens in a new tab)Google Playlisting (opens in a new tab)

Parivar AI · how it's built

Real-time backendPublic source

RideWave

The engine behind a ride-hailing app: it finds nearby drivers the moment you book and sends them the ride instantly.

1 query
GEORADIUS finds every nearby driver
0 broadcasts
offers go to addressed sockets only

Key technologies: Redis GEO · Socket.IO · MongoDB

View sourcefor RideWave (opens in a new tab)

RideWave · how it's built

Request path

  1. PassengerPOST /booking · JWT
  2. Express APIfare · booking saved as pending
  3. Redis GEOGEORADIUS → nearby drivers
  4. Socket.IOnewBooking to specific sockets
  5. Driver acceptsoffer withdrawn from the rest
MicroservicesPublic source

Airline Booking System

An airline booking backend split into five small services, so a slow email server never slows down a booking.

5 services
each owns its models and migrations
Async
a booking never waits on email

Key technologies: RabbitMQ · API gateway · MySQL

View sourcefor Airline Booking System (opens in a new tab)

Airline Booking System · how it's built

Request path

  1. Clientx-access-token
  2. API Gatewayrate limit · JWT verified by Auth
  3. Bookingseats · Sequelize on MySQL
  4. RabbitMQpublish, respond immediately
  5. Reminderconsume · Nodemailer
Code execution platformPublic source

AlgoForge

A LeetCode-style judge that safely runs other people's code in a sandbox and shows the verdict live.

256 MB · 2 s
memory and time cap per sandboxed run
4 languages
Python, Java, C++, JavaScript

Key technologies: BullMQ · Docker · Fastify · TypeScript

View sourcefor AlgoForge (opens in a new tab)

AlgoForge · how it's built

Request path

  1. Clientsubmit code
  2. SubmissionFastify · saved as pending
  3. BullMQSubmissionQueue on Redis
  4. Evaluatorfresh Docker container per run
  5. Socket.IOverdict pushed, no polling

Numbers here are properties of the code (service counts, sandbox limits, query shape), read from the source. None are traffic or user figures.

More systems

Experience

  1. Dec 2025 – Present

    Live on Google Play

    Parivar AI, Founder, Engineering

    • Founded and built an AI companion app where users chat with, and get voice calls from, AI family members in Hindi.
    • Built both the mobile app (Flutter) and the backend (Node.js) that runs the AI conversations, voice, and reminders.
    • Took it from zero to a public Google Play listing, then through review, launch, and post-launch fixes.
    parivarai.in(opens in a new tab)

  2. Mar 2026 – Present

    In development

    Ashvi Health, Founder, Engineering

    • Building a preventive health platform for Indian families.
    • Own the backend end to end: Node.js and TypeScript on AWS EC2 and S3, behind Nginx.
    • One API surface shared by the Flutter app and the Next.js web client.
    health-website-nine.vercel.app(opens in a new tab)

  3. Oct 2025 – Dec 2025

    Rablo, Backend Developer Intern

    • Built REST APIs in Node.js and MongoDB, cutting response times by roughly 30%.
    • Added secure login (JWT) and access control to every protected route.
    • Refactored a monolithic backend into controller and service layers, moving business logic out of route handlers.

  4. 2022 – 2026

    B.Tech in Computer Science & Engineering

    Ajay Kumar Garg Engineering College, Ghaziabad, India

How I build

  1. Latency is a feature.

    If the user does not need to wait for it, it goes on a queue. Emails, code runs, and AI calls happen in the background.

    in: Airline Booking · AlgoForge

  2. Design for failure.

    Servers fail and networks drop. One slow dependency should delay a task, not break the product.

    in: RabbitMQ between Booking and email

  3. Measure before optimising.

    Numbers decide what to fix. Guessing usually fixes the wrong thing.

    in: Rablo API response times

  4. Simple beats clever.

    A monolith where one process is enough, a service split where ownership demands it. Same for AI: fewer moving parts, checked outputs.

    in: RideWave, a monolith on purpose

AI engineering, on a backend foundation

The same rules as the backend work: behind a service boundary, measured, and honest about maturity. Here is where it has shipped and where it is going.

  1. Shipped

    AI in a live product

    Parivar AI talks to users in Hindi, calls them in a natural voice, and reacts to the photos they send.

    • LLM APIs
    • Prompt design
    • Conversation management
    • Text-to-speech
    • Image input

  2. Building

    AI grounded in your data

    Retrieval over your own documents, and structured outputs other services can consume.

    • RAG
    • Embeddings
    • Vector search
    • Structured outputs
    • FastAPI
    • Tool calling

  3. Next

    AI you can trust in production

    Agents with bounded autonomy, measured by evaluations and held in place by guardrails.

    • Agents
    • LangGraph
    • MCP
    • Evaluations
    • Guardrails
    • LLMOps
View full AI skill map
Production
Used in shipped products or public working systems
Hands-on
Actively building with it
Learning
Studying and experimenting

LLM applications

LLM features behind real APIs: typed inputs, typed outputs, streamed responses.

  • LLM API integration, Production
  • Prompt engineering, Production
  • Conversation management, Production
  • OpenAI APIs, Hands-on
  • Gemini APIs, Hands-on
  • Structured outputs · Pydantic, Hands-on
  • Function / tool calling, Hands-on
  • FastAPI services, Hands-on
  • Streaming · SSE, Hands-on
  • Async AI workflows, Hands-on

RAG systems

Retrieval pipelines that ground answers in sources and can prove it.

  • Embeddings, Hands-on
  • Semantic search, Hands-on
  • ChromaDB, Hands-on
  • LangChain, Hands-on
  • Retrieval pipelines, Hands-on
  • Metadata filtering, Learning
  • Hybrid search · BM25, Learning
  • Cross-encoder reranking, Learning
  • Citations · grounding, Learning
  • LlamaIndex, Learning
  • Pinecone, Learning
  • RAG evaluation · RAGAS, Learning

AI agents

Stateful, tool-using workflows with explicit control flow and bounded autonomy.

  • Tool-using agents, Hands-on
  • Memory · context management, Hands-on
  • LangGraph, Learning
  • Stateful workflows, Learning
  • MCP servers and tools, Learning
  • Multi-agent systems, Learning
  • CrewAI · AutoGen · smolagents, Learning

Multimodal & voice

Speech and vision as first-class inputs and outputs.

  • Text-to-speech, Production
  • Image input to LLMs, Production
  • Speech-to-text, Learning
  • Voice agents, Learning
  • CLIP · computer vision, Learning

Production AI

The backend half of AI: evaluation, guardrails, observability, cost, and latency.

  • Retries · fallbacks, Hands-on
  • Output validation, Hands-on
  • Latency · cost optimisation, Learning
  • Prompt caching, Learning
  • Evaluations · LLM-as-judge, Learning
  • Guardrails · prompt-injection defence, Learning
  • Observability (Langfuse, LangSmith), Learning
  • Local LLMs · Ollama, Learning
  • LLMOps, Learning

Deep learning foundations

Enough of the model side to reason about what the API is doing.

  • PyTorch · training loops, Learning
  • CNNs · batch normalisation, Learning
  • RNNs · LSTMs · GRUs, Learning
  • Transformer architecture, Learning
  • Tokenisation · BPE, Learning
  • Fine-tuning, Learning

Not claimed: ML research or training models at scale. The deep learning layer is there so I understand what the model is doing behind the API.

Skills, by evidence

Production work first. Hands-on work is labelled and set smaller; anything I'm only studying lives in the AI skill map.

Production
Used in shipped products or public working systems
Hands-on
Actively building with it

Backend

  • Node.js, Production, used in Every backend I have built
  • Express.js, Production, used in RideWave, Airline, AlgoForge, Rablo
  • Fastify, Production, used in AlgoForge Submission Service
  • REST API design, Production, used in All backend work
  • Microservices, Production, used in Airline Booking, AlgoForge
  • API gateway, Production, used in Airline API Gateway
  • JWT auth · bcrypt, Production, used in Auth Service, RideWave, Rablo
  • Socket.IO, Production, used in RideWave, AlgoForge Socket Service
  • Rate limiting, Production, used in Airline API Gateway
  • Schema validation (Zod), Production, used in AlgoForge Evaluator
  • FastAPI, Hands-on, used in AI engineering projects

Data

  • MongoDB · Mongoose, Production, used in RideWave, AlgoForge, Rablo
  • MySQL · Sequelize, Production, used in Airline Booking (five services)
  • Redis, Production, used in GEO in RideWave, queues and socket maps in AlgoForge
  • SQL, Production, used in Airline Booking
  • Vector databases (ChromaDB), Hands-on, used in RAG projects

Distributed systems

  • RabbitMQ, Production, used in Airline Booking → Reminder Service
  • BullMQ, Production, used in AlgoForge submission and evaluation queues
  • Event-driven messaging, Production, used in Airline Booking, AlgoForge
  • Async job pipelines, Production, used in AlgoForge
  • Geospatial queries, Production, used in RideWave (Redis GEORADIUS)
  • Sandboxed execution, Production, used in AlgoForge (Docker per submission)

AI engineering

  • LLM API integration, Production, used in Parivar AI
  • Prompt engineering, Production, used in Parivar AI personas
  • Conversation management, Production, used in Parivar AI
  • Text-to-speech, Production, used in Parivar AI voice calls
  • Image input to LLMs, Production, used in Parivar AI photo reactions
  • Embeddings, Hands-on
  • Retrieval pipelines, Hands-on
  • Structured outputs · Pydantic, Hands-on
  • Function / tool calling, Hands-on
  • FastAPI services, Hands-on

Infrastructure

  • Docker · dockerode, Production, used in AlgoForge Evaluator
  • Structured logging (Winston), Production, used in AlgoForge
  • Queue monitoring (Bull Board), Production, used in AlgoForge
  • Git · Postman, Production, used in All work
  • AWS EC2 · S3, Hands-on, used in Ashvi Health (in development)
  • Nginx, Hands-on, used in Ashvi Health (in development)
Product surface
Flutter · Next.js · React · Tailwind CSSShipped in Parivar AI, Ashvi Health, and this site
Languages
JavaScript · TypeScript · Python (hands-on) · C++

About

Portrait of Krrish Kumar
AKGEC · CSE 2022–2026 · Delhi NCR

I care about systems that work in the real world.

I started on the layer users never see: APIs, queues, and the wiring between services. A five-service airline backend taught me service boundaries and asynchronous messaging; RideWave and AlgoForge taught me real-time delivery and how to run untrusted work safely.

In December 2025 I founded Parivar AI and built both the Android app and the backend behind it, which is how I ended up shipping AI to real users. Next is going deeper on the AI side (retrieval, agents, evaluation) without leaving the backend behind.

Building
Parivar AILive on Google Play
Developing
Ashvi HealthPreventive health, in development
Studying
Distributed systemsMIT 6.824 · DDIA
Open to
Backend & AI rolesIncluding founding-engineer seats

Or ask the terminal

Everything on this page, from a command line. Output is generated from the same content files, so it cannot drift.

Let's build something that works at scale.

Looking for backend engineering, AI engineering, or founding-engineer opportunities. Email is fastest; I reply within a day.

Email me

krrishkumar2028@gmail.com

Based
Delhi NCR, India
Timezone
IST, UTC+5:30

Or leave a message

Krrish Kumar