Customer to architecture
Discovery, scope, tradeoffs, system design, rollout plan.
Forward deployed ownershipKaran Rajput
Forward deployed engineer and co-founder in Pune.
I turn unresolved customer problems into operated production systems.Inspect selected systems7 production web apps22 Indian languagesVoice + WhatsApp + Email + WebCloudflare + AzureIdea to rollout7 production web apps22 Indian languagesVoice + WhatsApp + Email + WebCloudflare + AzureIdea to rollout
Profile / operating principle
Built to run.
I work where product ambiguity meets production reality: learn the operation, choose the smallest reliable architecture, build the AI and application layers, then stay with the system after launch.
At CallMissed, that spans multilingual agents, seven web apps, a FastAPI platform, retrieval and memory, deployment automation, observability, and cost discipline.
04agent channels07production apps22Indian languages
From customer discovery through production operations, the work changes shape while ownership stays put.
Ownership does not move.
Agents to production
Voice, WhatsApp, email, web chat, tools, evals, memory.
Four live channelsCode to reliable service
React, Next.js, FastAPI, PostgreSQL, containers, observability.
Whole-stack deliveryCloud to calm operations
Azure, Cloudflare, DigitalOcean, Heroku, CI/CD, cost control.
Production ownershipField log
Current work
Proof under real constraints.
Co-founder & Forward Deployed Engineer at CallMissed
Take customer problems from discovery to production. Deploy frontier AI models (LiveKit voice, WhatsApp/Email/Web Chat agents) into real businesses across India - and own the whole stack underneath: architecture, infra, backend, frontend, rollout.
- 01Lead end-to-end deployments for design partners - scope, system design, build, rollout, and adoption - across an AI-native CRM that unifies WhatsApp, Voice, Email, and Web Chat behind 24/7 agents in 22 Indian languages.
- 02Architected and operate a production monorepo: 7 Next.js apps on Cloudflare Workers + a FastAPI backend on Azure Central India with PostgreSQL/pgvector for retrieval and agent memory.
- 03Built and shipped LiveKit voice agents and WhatsApp/Email/Web Chat AI agents - from prompt design and tool-use orchestration to evals and production rollout.
- 04Own CI/CD on GitHub Actions: preview environments, automated migrations, and zero-downtime deploys across web and API tiers; on-call and observability for the whole system.
Next.jsFastAPIAzureCloudflare WorkersPostgreSQL / pgvectorLiveKitAI AgentsGitHub Actions
Backend & React Intern at AnyFeast
Working on the consumer side of an AI-powered recipe-kit and meal-planning platform - building features that turn user taste preferences into personalised weekly meal plans and grocery orders.
- 01Shipping React components for the meal-planning and recipe-kit experience - focused on authentic local cuisines and fast, mobile-first interactions.
- 02Building and consuming backend APIs powering the AI meal planner, recipe catalog, and order flows for customers across India.
- 03Improving data wiring between the AI personalisation layer and the storefront so plans, ingredients, and substitutions stay in sync end-to-end.
- 04Collaborating across product, design, and ops to land features into production on a fast iteration cycle.
ReactNode.jsREST APIsAI Integrations
Selected systems / scroll film01 : 06
CallMissed.com
AI-native CRM monorepo deployed and operated in production: 7 Next.js apps on Cloudflare Workers (marketing, dashboard, docs, API playground, status, blog) plus a FastAPI backend on Azure Central India with PostgreSQL/pgvector, LiveKit voice agents, and GitHub Actions CI/CD. Unifies WhatsApp, Voice, Email, and Web Chat with 24/7 AI agents across 22 Indian languages.
Next.js 16FastAPIAzureCloudflareLiveKitWhatsApp APIMeta Tech Provider
Read case study ↗More shipped work
Products, infrastructure, automation, commerce, and applied computer vision.
View the full archive
Available for the next difficult build, in Pune or remote.
Bring thedifficult part.
Forward deployed, product engineering, and ambitious AI system work.
Start a conversation ↗