~ $ Senior Full-Stack AI Engineer_

I build AI products end to end.

LLM orchestration, distributed backends, and the cloud infrastructure under them.

Founding engineer at Aereo. Took the platform from zero to enterprise scale, and now lead 10 engineers.

  • Bengaluru, India
  • ·Remote-friendly
Book a callView workDownload résumé
Priyanshu Agarwal, Senior Full-Stack AI Engineer

TOP 3% TALENT

Vetted byHire me
0→1
Product built as founding engineer
Aereo platform, from nothing to multiple enterprise clients
10
Engineers led
Across web, mobile, and infrastructure platform pods
90%
AWS cost reduction
Redesigned distributed terrain-processing pipelines
50K+
Images per workflow
Production distributed photogrammetry systems
1
Granted patent
Geospatial data analytics architecture, Oct 2025

What I do

Five ways I'm usually brought in

Framed by the problem you're solving, not by the tools. Most engagements are some combination of these.

01

0→1 product delivery

See details →
01

0→1 product delivery

Taking a product from nothing to paying enterprise customers: frontend, backend, infrastructure, mobile launch, and the DevOps to keep it running. I've done this once end to end and then scaled it.

Product architectureFull-stackEnterprise rollout
02

AI systems that survive production

See details →
02

AI systems that survive production

RAG pipelines, LangGraph agent orchestration, deterministic inference orchestration, and context management for constrained-context LLM systems. Built to run reliably, not just demo well.

LangChainLangGraphRAGvLLMpgvector
03

Distributed backends & data platforms

See details →
03

Distributed backends & data platforms

Python, Django, and FastAPI services with async workflows, event-driven processing, and AWS Batch orchestration, designed for throughput at 50,000+ item scale without the cost curve to match.

PythonDjangoFastAPIAWS BatchPostgreSQL
04

Cloud infrastructure & deployment

See details →
04

Cloud infrastructure & deployment

Terraform from scratch, Kubernetes, hardened AWS networking, and CI/CD that holds up, including fully air-gapped on-premise deployments for enterprises that can't use public cloud.

TerraformKubernetesAWSDockerGitLab CI/CD
05

Engineering leadership

See details →
05

Engineering leadership

Leading 10 engineers across three pods: architecture and technical direction, cross-functional delivery, hiring and interviewing across SDE and DevOps pipelines, and structured onboarding and mentoring from SDE-1 to SDE-3.

Tech leadershipHiringMentoringDelivery

Live demo

The kind of system I build

A distributed geospatial platform: processing clusters, air-gapped enterprise deployments, and edge tile delivery. Click any node to see the services behind it.

Rendered with CesiumJS, the geospatial engine I've worked in for four years. Drag to rotate, scroll to zoom.

Selected work

Six things worth reading about

Each one is a real system with a real constraint and a number attached. Roughly two minutes each.

0→1Then multiple enterprise clients

Building a geospatial platform from zero

Joined as a founding engineering member, built the platform from scratch, and scaled it to multiple enterprise clients.

PythonDjangoReactCesiumJS
Read case study
Air-gappedFull platform, zero public cloud

Air-gapped enterprise deployment

Enterprise customers who could not touch the public cloud still needed the full platform. So it had to ship to them.

KubernetesDockerPythonDjango
Read case study
RAG + agentsRunning on customer on-prem GPUs

Production AI workflows on constrained hardware

LLM features that had to be reliable, explainable, and run on on-premise GPUs with no cloud inference.

LangChainLangGraphRAGvLLM
Read case study
90%AWS cost reduction

Cutting distributed processing costs by 90%

Terrain processing worked, but the AWS bill scaled with it and 50GB+ jobs were failing in production.

PythonAWS BatchS3Multiprocessing
Read case study
97%Faster environment provisioning

Infrastructure and CI/CD from scratch

No IaC, manual provisioning, and a deployment pipeline that could not keep up with the team.

TerraformAWSECSGitLab CI/CD
Read case study
10Engineers across 3 pods

Leading 10 engineers across three pods

For clients who need someone to run the team, not just write the code.

ArchitectureHiringMentoringDelivery
Read case study

Experience

Founding engineer → engineering lead

Four years at one company, from the first commits through enterprise scale, which means I've had to live with my own architecture decisions.

  1. Senior Software Development Engineer (SDE-3) · Aereo

    Jul 2024 - Present

    Lead engineering execution across the web platform, mobile, and infrastructure pods, covering 10 engineers plus a Data Science team, driving architecture, technical direction, and cross-functional delivery.

    • Led and shipped the company's first dedicated mobile product, from inception to production launch.
    • Architected a self-contained platform package for secure air-gapped enterprise deployments on Kubernetes-based on-premise infrastructure.
    • Built production AI assistant workflows using LangChain, LangGraph, and RAG pipelines over geospatial datasets.
    • Established structured onboarding through an internal learning platform; conducted interviews and mentored engineers across SDE-1 to SDE-3 and DevOps hiring pipelines.

  2. Software Development Engineer (SDE-2) · Aereo

    May 2022 - Jul 2024

    Scaled the platform to major enterprise customers while rebuilding the infrastructure and processing pipelines underneath it.

    • Delivered 10+ major features in 3 months for enterprise onboarding of a state-owned mining enterprise.
    • Reduced terrain-processing AWS costs by 90% by redesigning distributed synchronization pipelines.
    • Architected and deployed Terraform IaC from scratch covering ECS, RDS Multi-AZ, EC2, CloudFront, NAT Gateway, and Secrets Manager.
    • Improved deployment turnaround by 50% with autoscaled GitLab Runner infrastructure sustaining 10 to 15 parallel pipelines.

  3. Software Development Engineer (SDE-1), founding engineering member · Aereo

    Sep 2021 - May 2022

    Joined at the start and built the platform from scratch, establishing the architecture the team still works in.

    • Unblocked a major steel enterprise deployment with self-hosted geospatial delivery satisfying government data residency requirements.
    • Built terrain tile generation systems, COG generators, MBTiles extraction workflows, and vector tile servers.
    • Designed foundational event-driven AWS Batch job orchestration systems.
    • Architected foundational frontend systems adopted as team-wide engineering standards; published the team's first shared internal frontend module.

  4. Software Engineer · SalesHandy

    Jan 2021 - Aug 2021

    Backend APIs and distributed campaign management workflows for a high-traffic SaaS product.

    • Built backend APIs and distributed campaign workflows using NestJS, TypeORM, and PostgreSQL.
    • Optimized high-traffic SQL queries, reducing production database load and improving API responsiveness.
    • Introduced automated testing infrastructure with Jest, improving deployment reliability.

B.Sc (Hons) Computer Science · S.G.T.B Khalsa College, Delhi · Jul 2017 - Oct 2020

PatentGranted Oct 2025

A System For Geospatial Data Analytics And A Method Thereof

Co-invented the core backend architecture of a distributed geospatial analytics platform.

View patent →
AwardSep 2022

Spotlight Award, Aereo

Recognized for contributions to backend scalability, distributed systems reliability, and enterprise platform delivery.

Stack

What I work in

Languages
Python/TypeScript/JavaScript/SQL
Backend
Django/FastAPI/NestJS/REST APIs/Microservices
Distributed systems
AWS Batch/Async workflows/Distributed processing/Event-driven systems
Cloud & infrastructure
AWS/Kubernetes/Terraform/Docker/GitLab CI/CD
Databases
PostgreSQL/PostGIS/Redis/pgvector
AI / LLM systems
LangChain/LangGraph/RAG/vLLM/PyTorch
Frontend
React/TypeScript/Redux/CesiumJS

Available for contract and consulting engagements

Tell me what you're building.

Whether it's an AI product that needs to hold up in production, a backend that's outgrown its architecture, or a team that needs technical direction, send me the problem and I'll tell you honestly whether I'm the right fit.

Book a callEmail me
LinkedIn
Connect
Résumé
Download PDF