open to remote roles · ai backend · real-time systems

I build AI systems that survive production.

I'm Arin Nigam, a backend engineer at Cityflo. I pitched, prototyped, and shipped Komo AI — a multi-agent platform the whole company runs investigations on — and built the real-time core of SageAI, an AI support backend that deflects 80% of 36,000 weekly chats.

chat contacts handled per week
36K

SageAI

resolved by AI, no human needed
80%

SageAI

fewer human contacts per 1K rides
56%

SageAI

daily users, 30–40 investigations/day
80+

Komo AI

// experience

Where I've shipped

Production systems with real users and real numbers — plus remote, async open-source work.

Cityflo

Software Development EngineerJul 2025 – Present · Mumbai
  • Pitched, prototyped in a 6-hour hackathon, and shipped Komo AI to production — a multi-agent service platform now used company-wide by 80+ people across engineering, ops, support, and founders — cutting investigation time from 30–60 minutes to 2–3 minutes across 30–40 daily investigations.
  • Designed the agent framework itself: spec-driven agent definitions, per-agent MCP tool servers, USD-budget-bound execution, and bidirectional agent handoff. Shipped two agents — Tech Mazdoor (17 tools) and CX-Pulse (11 tools) — integrating Aurora, ClickHouse, GitHub, Datadog, Sentry, Airflow, Celery, Linear, and CleverTap.
  • Built core components of SageAI, Cityflo's AI customer support backend in Go — 36K chat contacts/week with an 80% AI deflection rate, cutting human-handled contacts per 1,000 rides by 56%. Owned the LLM integration, AI→human handoff orchestration, and ticket sync.
  • Replaced SageAI's Flowise layer with an in-house agent service — FastAPI + PydanticAI agents on a conversation graph exposing 20+ tools across ride, ticket, payment, and driver domains, a fallback-wrapped LLM cascade via LiteLLM, and a realtime voice agent (Gemini Live over LiveKit WebRTC, call recordings to S3).
  • Engineered SageAI's per-room concurrency layer for stateful AI conversations across horizontally-scaled ECS: goroutine-per-room with buffered channels, Redis SetNX locks for exactly-one agent spawn, and Redis Pub/Sub for cross-instance command delivery.
  • Own Cityflo's new B2C subsegment — Luxe premium cabs and Bus Rentals — end to end: zone-based dynamic pricing with direct ops control, an operator dashboard that lifted ops efficiency 40%, API latency cut from 1s to 200ms, and a direct customer booking flow now going live.
Go
Python
FastAPI
Claude SDK
Redis
PostgreSQL
AWS ECS Fargate

Walmart Global Tech

Software Engineer InternJan 2025 – Jun 2025 · Bengaluru
  • Built admin portal APIs across BFF, adapter, and service layers for Confido, Walmart's seller-risk governance platform — replacing manual Postman/DB operations for 40+ risk analysts with RBAC-gated workflows, dual-admin approval flows, and SSO integration.
  • Built a regression automation framework (Airflow DAGs, BigQuery validation, Kafka outcome consumer) covering 5 production risk strategies — cutting validation cycles from 6–8 hours to under 30 minutes, enabling daily runs that previously ran weekly.
Java
Spring Boot
GraphQL
MongoDB
Airflow
BigQuery
Kafka

Palisadoes Foundation · GSoC '24

Open Source ContributorMar 2024 – May 2024 · Remote
  • Contributed to Talawa in a fully-remote, async GSoC environment — designed Figma mockups, shipped GraphQL API endpoints, and led discussions in a community of 200+ developers.
Flutter
React
GraphQL
MongoDB
Docker

OpenLake

Open Source ContributorJan 2024 – Feb 2024 · Remote
  • Conceptualized and built an institute app serving 500+ users with 200+ daily actives; architected 10+ REST endpoints, improving server response times by 25%.
Flutter
Node.js
MongoDB

edu: B.E. — BITS Pilani · CGPA 8.1/10 · Jul 2021 – Jul 2025

// systems

Systems I've built

Three run in production at Cityflo today. The rest I designed, shipped, and operate myself.

Komo AI

production @ cityflo

Python · FastAPI · Anthropic Claude SDK · MCP · AWS ECS Fargate

Multi-agent service platform I pitched and took from a 6-hour hackathon prototype to company-wide production. Spec-driven agent definitions, per-agent MCP tool servers, USD-budget-bound execution, and bidirectional agent handoff across 28 integrated tools.

80+ daily users 30–40 investigations/day 30–60 min → 2–3 min per investigation

SageAI

production @ cityflo

Go · Gin · FastAPI · PydanticAI · LiteLLM · Redis Pub/Sub · PostgreSQL

AI support backend handling stateful conversations across horizontally-scaled ECS. I built the real-time concurrency core — exactly-one agent spawn via Redis SetNX, cross-instance Pub/Sub — and the in-house agent service that replaced Flowise: PydanticAI agents on a conversation graph calling 20+ backend tools, a fallback-wrapped model cascade, and a realtime voice agent over LiveKit WebRTC.

36K chats/week 80% AI deflection voice + chat 20+ tools

Luxe & Bus Rentals

production @ cityflo

Python · Django · React · PostgreSQL · zone-based pricing engine

Cityflo's new B2C subsegment, launched 2025 — I own the entire architecture end to end. Designed zone-based dynamic pricing with direct operational control, built the operator dashboard for vehicle and driver management, and shipped the booking flow now going live for direct customer bookings.

solo-owned, end to end APIs 1s → 200ms 40% ops efficiency gain

Your Health Copilot

live product

Next.js · LangChain · RAG · Supabase · AWS

Full-stack dashboard where 100+ users upload and query personal health reports. RAG pipeline with 75% higher accuracy on multi-document queries; similarity search tuned for 40% lower latency.

100+ users 40% faster retrieval

GTA-VI Resume Builder

live product

Three.js · LangChain · Gemini · S3 · Upstash

Interactive 3D resume builder in a GTA VI-inspired world — AI-driven resume generation rendered in an explorable scene. 100+ resumes created.

100+ resumes created

MySummary Pro

micro-saas

Node.js · Next.js · Gemini · Docker · AWS

WhatsApp AI chatbot that summarizes websites and YouTube links. Full SaaS platform with integrated analytics, handling 1,000+ monthly API calls.

1K+ API calls/month 90% summary efficiency

WebAppScaler

systems deep-dive

Node.js · Kafka · RabbitMQ · Redis · MongoDB · Docker

Load-bearing architecture study: queueing, caching, and messaging patterns pushed to 10,000+ concurrent requests with 40% faster API responses.

10K+ concurrent requests

more on github — Smart Insti (AutomateX winner, 200+ DAU), PlayPals (Quark runner-up), Socio, Cricstat → github.com/ArinNigam

// demos

Watch it run

Screen recordings of the products above, working. No slides — just the software.

cityflo luxe

Luxe operator dashboard

Vehicle, driver, and fare operations for the premium cab business — built solo, end to end.

cityflo rentals

Bus Rentals booking flow

Group commute bookings going direct-to-customer — no more manual ride confirmations.

health copilot

Your Health Copilot — demo

Upload personal health reports and query them through the RAG pipeline.

health copilot

Your Health Copilot — walkthrough

A deeper pass through the dashboard and multi-document queries.

automatex winner

Smart Insti app

The academic platform that won AutomateX 2024 — 200+ daily active users.

// stack

The stack I run

Grouped the way the systems are layered — from the model call down to the metal.

ai / llm
Anthropic Claude SDKMCPMulti-agent orchestrationPydanticAILiteLLMRAGVoice agents (Gemini Live)Vector embeddingsLangChain
backend
Go (Gin)Python (FastAPI)Java (Spring Boot)RESTGraphQLMicroservices
real-time
Redis Pub/SubGoroutines & channelsWebSocketsLiveKitWebRTCKafka
data & infra
PostgreSQLAuroraClickHouseMongoDBRedisBigQueryAirflow
cloud / ops
AWS ECS FargateS3SQSGCPDockerCI/CDDatadogSentry
languages
GoPythonJavaTypeScript

// signals

Competitive record

winnerAutomateX Hackathon 2024Smart Insti — Flutter + backend academic platform, 200+ daily active users
top 50Meesho Dice Challenge 2024Pitched a 3D AR prototype for in-app product visualization
runner-upQuark Hackathon 2024 · BITS GoaPlayPals — peer-to-peer activity matchmaking with Steam integration
contributorGoogle Summer of Code '24 · PalisadoesRemote open-source work in a community of 200+ developers
1000+Problems solvedLeetCode global rank under 30K

// attachments

The paper trail

Everything above is verifiable. Résumé, source code, and live products — pull whichever you need.

// contact

Let's build your AI product

Hiring for a backend or AI systems role? Tell me what you're building.

arin.nigam@gmail.com

Mumbai, India — remote-first by preference

IST (UTC+5:30) — comfortable overlap with EU all day and US mornings/evenings; async-first from open-source and distributed-team work

status: open to remote full-time roles
focus: AI backend · multi-agent systems · real-time infra
response_time: < 24h