K/ AVAILABLE FOR WORK
AVAILABLE FOR WORKDOWNLOAD RESUME

HEY, I'M KAUSHIK 👋

I build AI-powered systems that solve
real problems.

I combine AI engineering with backend expertise to build scalable systems, experiment with modern models, and share everything I learn.

EXPLORE MY WORK →VIEW RESUME ⇩
// building useful things
const kaushik = {
  role: "AI Engineer",
  focus: ["AI", "Backend"],
  current: "Building in public"
};

while (true) {
  build(); learn(); share();
}
LLM
DATA
SYSTEM
TECH I WORK WITHLocal AILLMsAI AgentsRAGTypeScriptGoPostgreSQLDocker
</>
8Projects
▤
14Knowledge Notes
△
8AI Experiments
▶
8Videos
☆
5+Years Experience

FEATURED PROJECTS

VIEW ALL PROJECTS
◉

AI Code Review Agent

AI Products

An agentic review pipeline that turns repository context into structured, actionable code feedback.

TypeScriptNext.jsPostgreSQLLLM APIs
↗ CASE STUDY◉ GITHUB
◎

Document QA System

AI Products

A retrieval system built to make document answers traceable, measurable, and fast enough to use.

PythonFastAPIPostgreSQLpgvector
↗ CASE STUDY◉ GITHUB
▥

Event Processing Core

Backend

A resilient backend for high-volume events with explicit retries, idempotency, and observability.

JavaKafkaRedisPostgreSQL
↗ CASE STUDY◉ GITHUB

LATEST KNOWLEDGE

VIEW ALL KNOWLEDGE

How Agent Tool Calling Works

AgentsTool Calling

Aug 28, 2026 • 7 min read

RAG Retrieval and Reranking

RAGRetrievalReranking

Aug 27, 2026 • 9 min read

Idempotency in Distributed Systems

Distributed SystemsReliabilityQueues

Aug 22, 2026 • 8 min read

A Mental Model for PostgreSQL Indexes

PostgreSQLIndexingQuery Planning

Aug 24, 2026 • 8 min read

AI LAB

VIEW ALL EXPERIMENTS

Experiments, benchmarks and things I'm exploring.

›_ EXP-014

RAG Retrieval on Local Models

RAG · Local AI · Benchmarks

Recall@5
91.4%
P95 latency
312ms
Queries
500

Aug 26→

›_ EXP-015

Agent Tools Under Failure

Agents · Evaluation · Reliability

Pass rate
86.4%
Unsafe retries
2.1%
Tests
120

Aug 21→

›_ EXP-016

Can LLMs Generate Schemas?

LLMs · Structured Output · Evaluation

First-pass valid
89.8%
After repair
97.5%
Runs
80

Aug 18→

VIDEOS

VIEW ALL VIDEOS

Building a Code Review Agent

▶22:15
▰   VISIT MY YOUTUBE CHANNEL →

Building a Code Review Agent

22:15

Deploying RAG Systems

18:07

Evaluating Agent Tools

12:21

LEARNING JOURNEY

VIEW ALL NOTES

My learning roadmap and notes.

ROADMAPNOTESBOOKSCOURSES

CURRENT FOCUS

  • ✓RAG evaluation
  • ✓Agent tool reliability
  • ✓Local inference
  • ✓KV cache
  • ○Agent memory
  • ○Model routing

RECENT NOTES

← BACK TO ALL ARTICLESSHARE   ◇   in   ↗

RAG · Evaluation   Aug 27, 2026   •   7 min read

What I Learned Building My First RAG Evaluator

The mistakes, metrics, and architecture changes that made evaluation useful instead of decorative.

retrieval.evaluate({faithfulness: 0.91,latency_ms: 312})
READ FULL ARTICLE   →