Shantanu Badmanji.

Backend Engineer | Infrastructure Optimizer | AI Systems Architect

Building scalable platforms and developer tooling on AWS. I architect AI-powered systems adopted by 50+ managers, optimize infrastructure (81% cost reduction), and ship production code that matters.

About Me

I'm a Backend Engineer with 2+ years of production experience architecting AI-powered platforms and developer tooling. At DevDynamics, I've grown from intern to SDE 2, shipping features used by thousands of engineering teams.

My expertise spans full-stack backend: multi-agent LLM systems, microservices architecture, database performance optimization at scale (60M+ rows), and cloud infrastructure optimization (reduced AWS spend by 81%).

I'm passionate about solving hard engineering problems—whether it's building AI copilots adopted by 50+ managers, mentoring junior engineers, or cutting infrastructure costs through targeted optimization.

Current Role

SDE 2 @ DevDynamics

Location

Pune, India

Education

BTech IT (2021-25)

VIIT Pune • CGPA 9.1/10

Featured Projects

Dyna AI

Multi-agent Copilot Backend

Production

Multi-agent copilot backend using LangGraph (supervisor/collector agents) and FastAPI with SSE streaming and parallelised data fetching. Adopted by 50+ managers across 100+ chat sessions.

Impact: 50+ manager adoption, 100+ active chat sessions

Tech Stack: FastAPI, LangGraph, PostgreSQL, AWS Lambda, SSE

Key Features: Multi-metric aggregation, parallel data fetch, streaming responses

FastAPILangGraphAI/LLMAWS

Refacto

AI-Powered PR Review Platform

Production

End-to-end AI-powered PR review platform built in TypeScript. Architected ephemeral AWS infrastructure with Lambda-based repo cloning ensuring client source code is never persisted beyond a single invocation.

Architecture: ECS, SQS, Lambda, S3 with ephemeral compute

Security: Zero-persistence code handling, ephemeral execution

Collaboration: Built directly with CTO from prototype to production

TypeScriptAWSServerlessAI

81% AWS Cost Reduction

Infrastructure Optimization & Architecture

$4.2K → $800/mo

Reduced AWS infrastructure spend from ~$4,200 to ~$800 per month through targeted compute and architecture optimization. Migrated ECS Fargate to self-hosted EC2, implemented peak/off-peak cron scaling, consolidated 10 load balancers into 2.

Optimizations: Fargate → EC2, ECS autoscaling, load balancer consolidation

Database: Cold-storage migration, replica removal, instance right-sizing

Result: 81% reduction while maintaining reliability

AWSECSAuroraOptimization

PostgreSQL Performance Tuning

Query Optimization at 60M+ Row Scale

>30s → <500ms

Diagnosed and resolved critical production correctness and performance issues in a 60M+ row PostgreSQL table. Applied CTE rewrites and index-friendly query restructuring to optimize 200+ generated query variants.

Issues Fixed: Timezone handling bug, 50+ shared SQL clauses

Performance: Worst-case query from >30s to <500ms, API latency 7–10s to 1–2s

Impact: Fixed data accuracy for international customers

PostgreSQLQuery OptimizationPerformance

Work Experience

Software Development Engineer 2

DevDynamics • June 2026 – Present

Pune, India
  • • Reduced AWS infrastructure spend by ~81% ($4,200 → $800/month) through targeted optimization
  • • Own CI/CD and production reliability across 10+ CodePipeline pipelines and 9 ECS services
  • • Mentored 2 engineers through onboarding to independent production ownership on AST-based codebase engine
  • • Authored onboarding guides and KT sessions to reduce ramp-up friction for future hires

Software Development Engineer 1

DevDynamics • June 2025 – May 2026

Pune, India
  • • Diagnosed and resolved critical production issues on 60M+ row PostgreSQL table, reducing query time from >30s to <500ms
  • • Owned Refacto (AI-powered PR review platform) end-to-end from prototype to production
  • • Architected Dyna AI (multi-agent copilot), adopted by 50+ managers across 100+ chat sessions
  • • Integrated Amazon Bedrock, extended VCS support (GitHub, GitLab, Azure, Jira), designed PostgreSQL advisory locks

Backend Development Intern

DevDynamics • June 2024 – May 2025 (Full-time, concurrent with BTech)

Pune, India
  • • Delivered 7+ third-party integrations (Asana, Azure DevOps, Bitbucket, Google Calendar, Slack) using webhook-driven patterns
  • • Built priority-based background data processing system in Python with dual-daemon architecture
  • • Engineered multi-user API token management system with rate-limit-aware rotation
  • • Developed appraisal automation (Sprint Reports, Contributor Reports) using GenAI pipelines, used by 80+ managers

Data Engineer Mentee

IBM SkillsBuild's Global Remote Mentorship Program • Oct 2023 – Feb 2024

Remote
  • • Designed and maintained incremental ETL pipelines ingesting daily CSV delta feeds from IBM Cloud Object Storage into IBM DB2
  • • Implemented transformation logic (data sanitisation, column manipulation, value formatting) via SQL scripts in IBM Query
  • • Implemented CDC (Change Data Capture) using hash-based change detection and auto-incremented surrogate keys
  • • Ensured downstream consumers always read from a consistent, fresh state via cron-scheduled orchestration

Technical Skills

Backend & Frameworks

NestJSExpressSpring BootFastAPINode.js

Databases

PostgreSQLAWS AuroraMySQLMongoDBIBM DB2

Cloud & DevOps

AWS ECSLambdaSQSS3RDSCodePipelineDockerLinux

Languages

TypeScriptPythonJavaGoJavaScriptSQL

Specialized Concepts

MicroservicesEvent-Driven ArchitectureQuery OptimizationDistributed SystemsAI/LLM SystemsRAGETL PipelinesCDCConcurrency ControlAST Analysis

Technical Writing

I write about cloud optimization, AI systems, database performance, and backend engineering

Read My Articles on Hashnode →

Let's Connect

Open to discussions about backend engineering, infrastructure optimization, and AI systems