Now · Senior Software DeveloperSenior SDE at Gramener

I engineer AI systems and platforms that hold up at national scale.

Senior full-stack engineer with 5+ years shipping production software, from schema to UI. Today I build Generative AI and analytics for India's National Health Authority across PM-JAY and ABDM.

agent.trace illustrative
❯
plandecompose into sub-tasksqueued
retrievepgvector · program docsqueued
querypostgres · aggregatequeued
reasonllm · synthesize findingsqueued
guardredact PII · cite sourcesqueued
renderdashboard + narrativequeued
RAG · agents · evals · guardrailshuman-in-the-loop
5+
years shipping production systems end-to-end
10M+
app downloads on a platform I helped build and run
National
scale: public health, citizen apps and live campaigns
IEEE
published research on GAN-generated image detection
01 / Capabilities

From an ambiguous problem to a production system, and everything in between.

I work across the whole stack, so the AI, the data, the APIs and the interface are designed together rather than stitched together at the end.

01

Generative AI & agentic systems

LLM applications that answer real questions over real data, with retrieval, tool use, guardrails and evaluation built in from day one.

  • RAG pipelines
  • Agentic workflows
  • pgvector
  • Prompt & eval design
  • Guardrails
02

Full-stack product engineering

High-traffic web and mobile platforms owned end-to-end, from database schema and REST APIs to fast, accessible, multilingual frontends.

  • React
  • Next.js
  • Angular
  • Node.js
  • Python
  • CMS
03

Data platforms & analytics

Turning large program datasets into dashboards and decision support that policy and operations teams actually use.

  • PostgreSQL
  • MongoDB
  • Kafka
  • RabbitMQ
  • Dashboards
04

Architecture & technical advisory

System design, build-vs-buy calls and delivery plans for teams that need to ship something serious without over-engineering it.

  • System design
  • AWS
  • Docker
  • Kubernetes
  • Terraform
  • CI/CD
02 / Selected work

Systems people depend on, not just demos.

A few engagements across public health, government and civic tech, and the enterprise platforms behind them.

2026 — nowGramenerPublic health · GenAICurrent

National Health Authority: PM-JAY & ABDM

Data, analytics and Generative AI for Ayushman Bharat PM-JAY and the Ayushman Bharat Digital Mission, India's national health assurance scheme and digital health stack.

  • LLM-powered applications, RAG pipelines and agentic workflows over large-scale health program data
  • Insight and decision-support tools for policy and operations teams
  • Full-stack applications, APIs and interactive dashboards
Stack
PythonLLMsRAGAgentsReactPostgreSQL
2023 — 2026OpalinaGovernment · Citizen apps

Narendra Modi website & mobile app

A secure, multilingual citizen platform with 10M+ app downloads, driven by a dynamic CMS.

  • Full-stack features across web and mobile surfaces
  • Multilingual content delivery with dynamic CMS integration
  • Security, responsiveness and reliability for a very large public audience
Stack
ReactAngularNode.jsCMS
2023 — 2026OpalinaLive national campaigns

Pariksha Pe Charcha & Mann Ki Baat

High-traffic national campaigns where participation spikes hard around live events.

  • Scalable, event-driven workflows for bursty traffic
  • Real-time data handling and performance optimization
Stack
Node.jsEvent-drivenReal-time
2023 — 2026OpalinaCivic tech · Canada

PushPolitics

A civic engagement platform connecting supporters with politicians across every level of Canadian government.

  • Led full-stack development and owned the system architecture
  • Responsive React/Angular frontend and Node.js REST APIs
  • NationBuilder integration
Stack
ReactAngularNode.jsNationBuilder
2023 — 2026OpalinaEnterprise · Media

Enterprise digital asset management platform

A production system for organizing, searching and governing large audio, video and photo libraries.

  • Database schema and access-control design
  • AI-assisted metadata generation
  • Docker-deployed, CDN-backed delivery pipeline on AWS
Stack
Next.jsPostgreSQLAWS EC2/RDSDockerLLMs
2021ResearchDeep learning

GAN-generated fake image detection ↗

A frequency-spectrum and deep-learning approach for detecting GAN-generated images. Experiments on ~1,000 StyleGAN2-ADA images reported accuracy above 99%. Published in IEEE.

Stack
PythonDeep learningStyleGAN2-ADA
03 / Approach

How I work.

The principles I bring to every engagement, whether it is a GenAI pilot or a platform serving millions.

/01

Start from the decision, not the model.

Who acts on the output, and what changes when they do? That defines the system, the data and the success metric.

/02

Evals before demos.

LLM features ship with test sets, guardrails and citations, so quality is measured rather than assumed.

/03

Boring infrastructure, sharp product.

Proven databases, queues and cloud primitives underneath, so the innovation budget goes where users feel it.

/04

Own it end-to-end.

Schema, API, UI and deploy pipeline in one head means fewer handoffs and faster, safer iteration.

“Public-scale software has no quiet launch. Design for the spike, the edge case and the person on a slow phone.”
— a lesson from national campaigns
04 / Experience

Experience.

Sep 2026 — nowGramenerSenior Software DeveloperNew Delhi+
  • Working with the National Health Authority (NHA) on Ayushman Bharat PM-JAY and the Ayushman Bharat Digital Mission (ABDM), building data and analytics solutions for India's national digital health ecosystem.
  • Building Generative AI solutions (LLM-powered applications, RAG pipelines and agentic workflows) that turn large-scale health program data into actionable insights for policy and operations teams.
  • Developing full-stack applications, APIs and interactive dashboards with Python, React and PostgreSQL.
May 2023 — Aug 2026Opalina TechnologiesSoftware EngineerNoida+
  • Led full-stack development of PushPolitics, a Canadian civic engagement platform, owning system architecture, React/Angular frontend, Node.js REST APIs and NationBuilder integration.
  • Contributed to high-traffic national campaigns, Pariksha Pe Charcha and Mann Ki Baat, building scalable, event-driven workflows with real-time data handling.
  • Developed and maintained full-stack features for the Narendra Modi website and mobile app (10M+ downloads), with secure, multilingual, CMS-driven experiences.
  • Designed and built an enterprise digital asset management platform on Next.js, PostgreSQL and AWS with AI-assisted metadata generation and a CDN-backed delivery pipeline.
  • Delivered multiple CMS-powered microsites using reusable responsive components.
Feb 2021 — May 2023DXC TechnologySoftware EngineerNoida+
  • Delivered end-to-end software solutions with cross-functional engineering and product teams.
  • Built responsive web interfaces, backend services, databases and reusable REST APIs focused on performance, reliability and maintainability.
  • Tested, debugged and upgraded production systems in Agile delivery environments using Jira and Confluence.
05 / Stack

Tools I reach for.

AI / ML

  • LLMs
  • RAG
  • Agentic AI
  • Prompt engineering
  • pgvector
  • Deep learning

Languages

  • TypeScript
  • JavaScript
  • Python
  • SQL
  • Java

Frontend

  • React
  • Next.js
  • Angular
  • HTML/CSS
  • Accessible UI

Backend

  • Node.js
  • Express
  • FastAPI
  • Django
  • Flask
  • Kafka
  • RabbitMQ

Data

  • PostgreSQL
  • MongoDB
  • Drizzle ORM
  • Vector DBs

Cloud & delivery

  • AWS
  • Docker
  • Kubernetes
  • Terraform
  • Vercel
  • CI/CD
  • Jest
06 / Background

Education & research.

Education · 2017 — 2021

B.Tech, Computer Science & Engineering

Amity University, Uttar Pradesh · CGPA 7.57 / 10

Research & side projects

IEEE paper on GAN image detection; real-time CCTV surveillance on Raspberry Pi

An ensemble-learning model with OpenCV running on a Raspberry Pi camera for real-time surveillance.

Read the paper ↗
07 / Contact

Have a hard problem worth solving?

Open to senior engineering roles, GenAI consulting and collaborations. Tell me what you are building and I will reply within a couple of days.