PERSONAL INTERNET OBJECT / 2026 / MUMBAI

advika
kharat.

[SDE I @ AMAZON]

I build backend systems, cloud infrastructure, and AI-powered products. Sometimes I make tiny internet things just because I think they're fun.

advika.exe—_ □ ×
C:\USERS\ADVIKA> whoami
advika_kharat
software engineer
amazon / aws
mumbai, india
C:\USERS\ADVIKA> cat interests.txt
systems
databases
artificial intelligence
weird websites
music
coffee coffee coffee
C:\USERS\ADVIKA> █
note to self:
make cool things.
* probably
overengineering this
   .------.
  /  o  o  \
 |     ∆     |
  \  _____  /
   '-------'
BACKEND✦AWS✦POSTGRESQL✦AI✦SYSTEMS✦DATABASES✦COFFEE✦MUSIC✦BACKEND✦AWS✦POSTGRESQL✦AI✦
01
CURRENTLY / PREVIOUSLY

work.log

[ EXPERIENCE ]
01
AUG 2025 — NOW

SOFTWARE DEVELOPMENT ENGINEER I

AMAZON
  • Strengthened environment isolation for four service tenants spanning 28 AWS Regions by designing and implementing Gamma/Production account separation.
  • Led deprecation of a legacy data-replication system across 100+ hosts and clusters in two cloud partitions, migrating workloads from DynamoDB Streams to PostgreSQL and reducing related on-call tickets by 60%.
  • Remediated 15+ production security findings, including WebSocket Origin validation, secure management of CSRF-signing secrets, sensitive-data redaction, and standardized security headers.
  • Drove service onboarding across two AWS Regions, navigating undocumented deployment paths, resolving cross-service dependencies, and reducing subsequent onboarding time from ~4 weeks to 2.
02
JAN 2025 — JUN 2025

SOFTWARE DEVELOPMENT ENGINEER INTERN

AMAZON
  • Built and deployed a GenAI-powered operations automation system for the RDS platform, achieving 78% ticket-resolution accuracy and reducing manual investigation effort and response time by 75%.
  • Developed production REST APIs in Core Java with 95% unit test coverage using JUnit and Mockito.
↑yes, I really do enjoy reading infrastructure logs.
02
THINGS I MADE

projects/

[ OPEN SOURCE-ish ]
01FLAGSHIP / SYSTEMS

QueryPilot

AI PostgreSQL Performance Analyzer
⌘

An AI-powered PostgreSQL performance lab that turns natural-language requests into SQL, validates queries, inspects real execution plans, and benchmarks index recommendations before accepting them.

›one verified optimization → >99% faster
PythonFastAPIPostgreSQLSQLGlotOllamaQwen 3
EXPLAIN
→ ME
YOUR
QUERY
02FLAGSHIP / PERSONAL

Mellow

AI Spotify Listening Intelligence
♫

A personal music intelligence system that lets you interrogate your Spotify history in natural language. Multi-stage AI planning turns questions into deterministic PostgreSQL queries and evidence-backed answers.

›your listening history, but actually searchable
Next.jsTypeScriptPostgreSQLSpotify APIOllama
03ARCHIVE / OLDER

Learnify

Learning Plan & Progress Platform
□

An earlier project for creating learner profiles, generating personalized learning plans, saving goals, and tracking progress through a simple web application.

›from the archive
FlaskPythonSQLAlchemyOpenAI APIBootstrap
03
A LITTLE CONTEXT

about.txt

[ READ ME ]

I'm a software engineer who likes understanding how things work underneath the abstraction.

Currently building cloud infrastructure at Amazon. Outside work, I tend to build small systems that solve oddly specific problems, experiment with AI, or make websites that look like they escaped from the early internet.

I like backend engineering, databases, distributed systems, developer tooling, and making complicated things feel simple.

ABOUT.DAT1.7 KB
LOCATIONMumbai, IN
EDUCATIONB.Tech Computer Engineering
COLLEGESPIT
CGPA8.4
STATUSOPEN TO REMOTE / RELOCATION
04
TOOLS / LANGUAGES / OBSESSIONS

toolbox/

Languages

PythonJavaTypeScriptJavaScriptSQL

Backend / Web

FastAPISpring BootNode.jsNext.jsReact

Cloud / Data

AWSPostgreSQLDynamoDBLambdaAPI GatewaySQSCloudFormationIAMCDK

AI / Tooling

OllamaAmazon BedrockSQLGlotGitBashCI/CD
05 / CONTACT

want to build
something interesting?

Email is probably the least weird way to reach me.