Skip to main content
Available for select roles

Abhishek Reddy Boddu

Python Backend & Applied AI Engineer

Former Amazon SDE I building production-grade FastAPI backends, orchestrating multi-agent workflows with LangGraph, deploying high-accuracy RAG vectors, and designing serverless AWS pipelines.

AWS & PIPELINES

Serverless ETL, Lambdas, SQS/SNS, S3 data handoffs, Step Functions.

APPLIED AI SYSTEM

LangGraph DAG topologies, pgvector similarity query, guardrails, evaluations.

01 / Python & FastAPI Backends

Building type-safe API gateways, high-throughput workers, pgvector indexes, and structured database layouts.

02 / Applied AI topologies

Deploying multi-agent LangGraph workflows, high-precision RAG context retrieval, prompt safety shields, and human-in-the-loop validation.

03 / AWS Cloud Infrastructure

Orchestrating serverless Lambda ETL jobs, Step Functions state machines, SQS queues, and S3-based handoff pipelines.

01 / Profile

Engineering dashboard & impact stats

Mindset

Amazon backend scale meets practical AI automation.

I'm Abhishek (Abhi). My focus is writing clean, predictable Python backend logic, building robust microservices, and integrating agent workflows where they actually solve user issues.

At Amazon, I learned that correctness, instrumentation, and failing gracefully are far more important than bleeding-edge buzzwords. Recently, I've translated that mindset to LangGraph workflow graphs, security guardrails, and RAG databases.

Core Stack

Focused Toolkit

Tools I use to build scalable APIs and AI pipelines.

PythonFastAPILangGraphLangChainPostgreSQLpgvectorReactNext.jsTypeScriptAWSDockerRedis
Backend Core: PythonCloud: AWS
Telemetry

Amazon Production Work

1.5M+

HR Request APIs

React, FastAPI, and AWS-backed services for HR requests.

80%

Self-Provisioning

Time-capture self-service kiosks across tablet/mobile platform.

60%

Lambda ETLs

Python ETL pipelines using Step Functions and S3 handoffs.

Open Source

GitHub Activity

View Profile →
Live activity synced directly from GitHub profile.

02 / Capabilities

What I can ship

The center of gravity is backend reliability with modern product surfaces and AI workflows where they actually help.

01

Backend APIs

FastAPI services, database-backed workflows, clean service boundaries, validation, migrations, and observable request paths.

02

Applied AI Workflows

LangGraph agents with RAG, tool calling, durable memory, evals, streaming UX, and human approval gates around sensitive actions.

03

Production Readiness

Failure-first engineering: retries, idempotency, guardrails, rate limits, security headers, SEO, and deployment-friendly structure.

03 / Projects

Work that proves the engineering

A tighter project set: one flagship AI/backend system, one live production-facing site, and one exploratory agentic workflow.

Flagship applied AI/backend project

AbhiMart AI Customer Support Agent

Production-style support backend using FastAPI, LangGraph, PostgreSQL/pgvector, Gemini, RAG, SSE streaming, durable memory, evals, observability, guardrails, and HITL refund approval.

Open

Tool-based order/product lookup with RAG over policy documents

LangSmith/local evals, OpenTelemetry/Jaeger, and Prometheus-style metrics

Prompt injection, PII, cross-customer access, and refund safety guardrails

FastAPILangGraphPostgreSQLpgvectorGeminiLangSmith

Live client-style Next.js project

aicloudhub.com

Migrated and built a production-facing AI consulting site with Next.js App Router, Sanity CMS, ISR, Vercel deployment, Redis rate limiting, Zod validation, SEO metadata, sitemap generation, and security headers.

Open

Improved Lighthouse/page-load latency by about 30%

Dynamic CMS routes with ISR instead of fully dynamic rendering

Contact/application protection through validation and rate limiting

Next.jsReactSanityVercelRedisZod

Agentic workflow prototype

Personal OS

Conversational life-management assistant exploring multi-intent routing, parallel LangGraph fan-out, typed UI messages, and interrupt-driven confirmation before database writes.

Open

Natural-language finance, journal, and media logging

Human confirmation before higher-risk writes

React UI cards rendered from graph outputs

LangGraphLangChainFastAPIReactSQLite

Honest scope: AbhiMart and Personal OS are hands-on production-style portfolio projects, not company production deployments. Real production experience comes from Amazon and aicloudhub is live client-style work.

04 / FAQ

Frequently Asked Questions

Quick insights into my backend engineering philosophy, agent workflows, and cloud patterns.