TECHNICAL PRODUCT MANAGER · HANDS-ON ENGINEER · FARIDABAD

I build production platforms, and the guardrails that let AI touch them safely.

Hands-on owner, lead and architect of RVM CAD’s four platforms and the AI guardrails around them. Former Air India A320 engineer. I verify, then ship.

Backend, full-stack web, DevOps, cloud, integrations, data and applied AI — owned end to end.

Open to remote, hybrid and on-site roles, in any location.

ENTER THE HANGAR · 3D PROTOTYPE

01 / CAPABILITIES

What I can own, end to end.

Every rating links to the systems that prove it. Green is in production; blue is from my roles; outlined is in training.

Backend & APIsIn productionContract-first APIs and services, tested in CI116 → 534API routes on the Scheduler, 30 Apr → 28 Oct 2026TypeScriptNode.jsNestJSFastAPIFull-stack web, end to endIn productionFrom the domain model to the deployed product, solo129commits on StrikeHike main, all mine, 14 Nov 2026TypeScriptNext.jsNode.jsPostgreSQLDevOps & release engineeringIn productionTests before deploy, probes after, rollback ready24production deploys on Team RVM, 16–27 Oct 2026, all greenGitHub ActionsDockerCaddyNginxCloudIn productionAWS, Azure, Google Cloud and Supabase in productionAWSLightsail runs StrikeHike and this siteAWSAzureGoogle CloudSupabaseAWSAWS Glue (in my roles)AzureAzure Data Factory (in my roles)AzureAzure AI Foundry (in my roles)InfrastructureIn productionReverse proxies, TLS, containers, process managersCaddyTLS and routing in front of StrikeHikeCaddyNginxPM2DockerIntegrationsIn productionPayments, messaging, AI connectors, data APIsRazorpayorders, webhooks, refunds and GST invoices on the Job PortalRazorpayMCPOpenAPIResendBackends for Flutter appsIn productionThe API contract and backend a mobile team builds on155operations in the Team RVM contract the Flutter app callsFlutter backendsOpenAPINode.jsPostgreSQLApplied AI & MLIn productionLLM features with guardrails, MCP connectors, MLOps11read-only tools in the Job Portal AI connectorPythonGeminiClaudeMCPData pipelines & schedulingIn productionIngestion and scheduled jobs that fail loudlyWeeklyevents bundle, Sunday 21:30 UTC on StrikeHikePythonGitHub ActionsAirflow (in my roles)AWS Glue (in my roles)Security by designIn productionLeast privilege, encryption and audit built inAES-256-GCMCHECK-enforced encryption in the AI trace storePostgreSQLSupabaseMCPOpenAPI
3,048
commits in 12 months 1: source for commits in 12 monthsmeasured
1,056
merged pull requests 2: source for merged pull requestsmeasured
5
deployed systems whose first commit is mine 3: source for deployed systems whose first commit is minemeasured
2.9
years in software since Jan 2024 4: source for years in software since Jan 2024measured

IN TRAINING

2 more await a source before they are shown. No certificate or course is claimed until one exists.

CAPABILITY / 01

Backend & APIs

In production

Production backends with tested routes, versioned migrations and an API contract that gates the deploy.

116 → 534

API routes on the Scheduler, 30 Apr → 28 Oct 2026

Source

155

operations in the Team RVM OpenAPI 3.1 contract

Source

2,556

Scheduler backend tests passing in CI, 28 Oct 2026

Source

WHERE I'VE DONE IT

STACK

TypeScriptNode.jsNestJSFastAPIPostgreSQLOpenAPI

ENGINE · CORE SYSTEM

CAPABILITY / 02

Full-stack web, end to end

In production

Whole products, not fragments: schema, API, interface, tests and the deploy that carries them.

129

commits on StrikeHike main, all mine, 14 Nov 2026

Source

94 / 95

commits on Singh Furncraft, built solo

Source

399 / 856

Job Portal commits on main, the most of anyone

Source

WHERE I'VE DONE IT

STACK

TypeScriptNext.jsNode.jsPostgreSQLSupabasePlaywright

ENGINE · FAN SYSTEM

CAPABILITY / 03

DevOps & release engineering

In production

Release paths that fail safely: tests gate the build, probes check the live system, and the previous images stay ready.

24

production deploys on Team RVM, 16–27 Oct 2026, all green

Source

Probes

health, data, TLS and a real API response after cutover, then rollback on failure

Source

2,556

Scheduler backend tests gating the deploy

Source

WHERE I'VE DONE IT

STACK

GitHub ActionsDockerCaddyNginxPM2Prometheus (in my roles)Grafana (in my roles)

ENGINE · OIL SYSTEM

CAPABILITY / 04

Cloud

In production

Production cloud: AWS Lightsail, Azure App Service on StrikeHike v1, Google Cloud with OAuth and the Meet/Calendar APIs, Supabase Postgres.

AWS

Lightsail runs StrikeHike and this site

Source

Azure

App Service ran StrikeHike v1

Source

GCP

Google Cloud project, OAuth and the Meet REST API (v2); Calendar API

Source

WHERE I'VE DONE IT

STACK

AWSAzureGoogle CloudSupabaseDockerCaddyAWS ▸AWS Glue (in my roles)Azure ▸Azure Data Factory (in my roles)Azure ▸Azure AI Foundry (in my roles)

ENGINE · PYLON SYSTEM

CAPABILITY / 05

Infrastructure

In production

The unglamorous layer that keeps a product reachable: proxies, certificates, containers and supervised processes.

Caddy

TLS and routing in front of StrikeHike

Source

Nginx · PM2

reverse proxy and process supervision on the Scheduler

Source

Docker

Compose stacks built and cut over on deploy

Source

WHERE I'VE DONE IT

STACK

CaddyNginxPM2Docker

ENGINE · MOUNTS SYSTEM

CAPABILITY / 06

Integrations

In production

Third-party systems wired in with signatures and an audit trail: payments, messaging, AI tool access and data APIs.

Razorpay

orders, webhooks, refunds and GST invoices on the Job Portal

Source

11

read-only tools in the Job Portal MCP connector

Source

57

read-only tools in the Scheduler ops/HR connector

Source

WHERE I'VE DONE IT

STACK

RazorpayMCPOpenAPIResend

ENGINE · GEARBOX SYSTEM

CAPABILITY / 07

Backends for Flutter apps

In production

Backends and contracts for Flutter apps: the mobile developer writes the UI; I own the API, data and release behind it.

155

operations in the Team RVM contract the Flutter app calls

Source

1,320

contract tests checking every reply

Source

Flutter

worker app built by teammates on the BadaBhai backend

Source

WHERE I'VE DONE IT

STACK

Flutter backendsOpenAPINode.jsPostgreSQL

ENGINE · IDG SYSTEM

CAPABILITY / 08

Applied AI & ML

In production

AI that touches real systems through a scoped identity, with traces and an audit trail around it.

11

read-only tools in the Job Portal AI connector

Source

57

read-only tools in the Scheduler ops/HR connector

Source

AES-256-GCM

CHECK-enforced encryption in the AI trace store

Source

WHERE I'VE DONE IT

STACK

PythonGeminiClaudeMCPAzure ▸Azure AI Foundry (in my roles)Airflow (in my roles)

ENGINE · FADEC SYSTEM

CAPABILITY / 09

Data pipelines & scheduling

In production

Data that arrives on time and jobs that say when they do not: scheduled ingestion, reconciliation and backups.

Weekly

events bundle, Sunday 21:30 UTC on StrikeHike

Source

10 min

in-process reconciliation with a soft lock on the Job Portal

Source

Gated

events bundle: quality gate, health check, rollback to the previous bundle

Source

WHERE I'VE DONE IT

STACK

PythonGitHub ActionsAirflow (in my roles)AWS ▸AWS Glue (in my roles)Azure ▸Azure Data Factory (in my roles)PostgreSQL

ENGINE · FUEL SYSTEM

CAPABILITY / 10

Security by design

In production

Access that is scoped, tokens that rotate, data encrypted at rest and sensitive calls on the record.

AES-256-GCM

CHECK-enforced encryption in the AI trace store

Source

Rotating

hashed refresh tokens; reuse revokes the family

Source

5 tables

append-only audit trail on the Scheduler

Source

WHERE I'VE DONE IT

STACK

PostgreSQLSupabaseMCPOpenAPI

ENGINE · FIREWALL SYSTEM

02 / SYSTEMS

Four platforms. One lead.

All work

01

BadaBhai

An AI placement platform for blue- and grey-collar India.

840 of 1,625 merged PRs are mine

HINGLISH PROFILING · ONE-PAGE RÉSUMÉS · PAYER TENANCY · PRIVACY RELAY

View system →

02

RVM CAD Job Portal

An industrial hiring platform: verified employer directory, job discovery, candidate applications, employer dashboards and hiring analytics.

399 of 856 commits on main, the most of anyone

JOB DISCOVERY · VERIFIED EMPLOYERS · CANDIDATE DASHBOARD · APPLICATIONS

View live ↗

03

RVM CAD Scheduler

The operations platform of a four-branch CAD/CAM training institute, from admissions and fees to staff attendance, HR and inventory.

116 → 534 API routes, 30 Apr → 28 Oct 2026

SCHEDULING · ATTENDANCE · GPS CHECK-INS · FACULTY

View system →

04

Team RVM

RVM CAD’s employee self-service Android app: attendance with location, requests and approvals, an inbox, a staff wall, a helpdesk, documents, an ID card and My Team.

24 production deploys, 16–27 Oct 2026, all green

ATTENDANCE · REQUESTS & APPROVALS · INBOX · STAFF WALL

View system →

I own, lead and architect all four; the engineers on them work under me, and each project page credits what they built. Figures from git, the GitHub API and CI, 28 Oct 2026. See the AI connectors

03 / INDEPENDENT BUILD

StrikeHike: built and run solo

Case study

A live research site that sets the US macro record beside the planetary sky, date by date: recession windows since 1854, a macro and volatility index, market series, and past events and upcoming dates with similar planetary geometry. I build and run it solo, from the domain to the deploy: a FastAPI backend and a Next.js 16 frontend in Docker Compose behind Caddy on AWS Lightsail. Every push to main is tested before the new stack goes up, then probed for health, TLS and a real API response; a failed probe restores the previous images and fails the deploy. Research, not investment advice.

Read the case study Architecture on Projectsstrikehike.com

StrikeHike Regime view: the Core Index macro atlas with the S&P 500 plotted and NBER recession windows shaded since 1854
StrikeHike Regime · the −50 to +50 macro atlas, with NBER recession windows since 1854

04 / SELECTED WORK

Evidence, not adjectives.

All work
Built and testedRVM CAD Job Portal and RVM CAD Scheduler

Both connectors are merged with full test suites. Production activation and the CEO acceptance test had not been verified when this was written (20 Oct 2026).

Doors, not keys: read-only AI access to production data

Two connectors that let an AI assistant answer operational questions from live systems — without a database credential, without write access, and without ever returning a half-right number.

11 read-only tools on the job portal connector

ProductionBadaBhai

Shipped in August 2026; the migration was applied to production with the release. Six AI call sites write to it.

An AI trace store that treats its own data as dirty

Recording every prompt and completion on an AI hiring platform, so bad answers can be investigated — encrypted at rest, erasable per worker, and built so raw text cannot reach the recorder by accident.

6 AI call sites traced

ProductionBadaBhai

Live since late August 2026 and refined through several rounds since. Measurements below are on the fixture set, not every live profile.

The one-page contract: résumés that fail closed

A résumé generator for workers who have never had one. The sheet is printed and handed to a supervisor, so "it rendered" is not good enough — it has to fit one page, use the right fonts and never invent a fact.

56 / 56 fixture sheets on exactly one A4 page

Built and testedBadaBhai

All three phases merged with tests in October 2026. Switching production to organisation tenancy is a controlled rollout step owned by the business; its current mode was not verified when this was written.

Multi-tenancy without a migration

Invited teammates on an employer account saw zero jobs, zero credits and zero candidates. The fix re-scoped fifteen business tables to the organisation — with no new column, no backfill and no change to the API.

15 business tables moved to organisation tenancy

Built and testedRVM CAD Job Portal

Merged with 616 of 616 tests passing at 48c3cab1 (23 Jun 2026). RVM approved describing the behaviour on 15 Nov 2026; production rollout state was not independently verified.

Interviews that schedule, record and file themselves

One interview, end to end in the Job Portal: a Google Meet space and invites, a consent gate before recording, a recording state machine, and Drive filing with metadata. Built in one commit, with 616 of 616 tests passing at that commit.

70 files in the feature commit

05 / RECENTLY BUILT

Recently built

Projects with meaningful changes in my recent GitHub activity, newest first. Private repositories are named, never linked.

Refreshed weekly from my GitHub activity · last changed

  1. StrikeHike

    Updated

    A research site that sets the US macro record beside the planetary sky, date by date: recession windows since 1854, a macro and volatility index, market series, and past events and upcoming dates with similar planetary geometry.

    • Next.js
    • React
    • Tailwind CSS
    • three.js
    • React Three Fiber
    Live site: StrikeHike (opens in a new tab)
  2. strikehike-growth-ops

    Updated

    Python codebase, created Nov 2026.

    • Python
    Private repository
  3. manifest_extension

    Updated

    AI Job Copilot: a Chrome extension (Manifest V3, React, TypeScript) that drafts a tailored cover letter and résumé PDF for a Greenhouse job application and fills the form for the user to review. FastAPI on AWS Lightsail, Supabase, Claude API. In development; not yet launched.

    • Claude API
    • FastAPI
    • PLpgSQL
    • Pydantic
    • pytest
    Private repository
  4. thisiskp.in

    Updated

    My portfolio and case-study site.

    • Astro
    • three.js
    • TypeScript
    • Vitest
    • Zod
    Live site: thisiskp.in
  5. quant_trading_app

    Updated

    Options analytics prototype for Indian markets: Greeks, implied volatility, open interest and put-call ratio for indices and F&O stocks, plus an option payoff calculator. FastAPI + Redis worker, React/Vite, Supabase. Demo data; information only, not investment advice.

    • FastAPI
    • JavaScript
    • PLpgSQL
    • Pydantic
    • pytest
    Private repository
  6. Kubernetes

    Upstream contribution · Updated

    A documentation fix merged into Kubernetes, kubernetes/kubernetes#142729 (cloud-provider, October 2026).

06 / LOGBOOK

How I got here

The longer version
  1. 2015 – 2017M.Tech, Cryogenics and Vacuum Technology, NIT Rourkela
  2. Feb 2019 – Jan 2024Aircraft Maintenance Engineer (DGCA Category A AME), Air India
  3. Jan 2024 – Aug 2025AIML Engineer, Boven Technologies
  4. Aug 2025 – Apr 2026ML-Ops Trainer, Futurense Technologies
  5. Apr 2026 – presentTechnical Product Manager (official designation), RVM Beyond Private Limited (RVM CAD)

07 / COMMON QUESTIONS

Quick answers

Who is Prakash Kantumutchu?

Prakash Kantumutchu (KP) is a technical product manager and hands-on engineer based in Faridabad, Delhi NCR, India. He owns, leads and architects RVM CAD’s four platforms and writes about safe AI access, AI reliability and verification.

What does he do at RVM CAD?

He is the official Technical Product Manager. In practice he owns, leads and architects BadaBhai, the RVM CAD Job Portal, the RVM CAD Scheduler and Team RVM, and the engineers on them work under him — architecture, backend and data, AI integration, CI/CD and production support.

What is he open to?

Remote, hybrid and on-site roles, in any location. The role families he fits are lead/senior platform engineer, AI systems or AI platform engineer, forward-deployed or solutions engineer, and technical lead.

What is StrikeHike?

StrikeHike (strikehike.com) is a live research site that sets the US macro record beside the planetary sky, date by date: recession windows since 1854, a macro and volatility index, market series, and past events and upcoming dates with similar planetary geometry. He builds and runs it solo. It is research, not investment advice.

Where is he based?

Faridabad, Delhi NCR, India.

What is his aviation background?

From February 2019 to January 2024 he was an Aircraft Maintenance Engineer (DGCA Category A AME) at Air India, working on Airbus A320-family fly-by-wire and powerplant systems under DGCA and EASA requirements.

How can I contact him?

Email [email protected], or use LinkedIn or GitHub. The Hire page also lists the role families he fits.

What can KP own end to end?

Backend and APIs, full-stack web, DevOps and release, cloud, infrastructure, integrations, mobile backends, applied AI, data pipelines and security by design. The capability board links every rating to the systems that prove it.

How do the capability ratings work?

There are three levels: in production, with a proof and a linked artefact; in my roles, from his own statement with the role and period; and in training, which needs a source. The board and /capabilities show the same data.

08 / OPEN CHANNEL

Get in touch

Hiring for an engineering role

Role families I fit, the scope of my current work, and how to reach me.

For hiring managers

Consulting or training

Short, scoped engagements on AI access, AI reliability and production audits, and training for engineers.

DIRECT: [email protected]

ESC

↑ ↓ to move · Enter to open · Esc to close