Software Engineering Portfolio

CHUNDURI
KAUSHIK
SOFTWARE ENGINEER
•
AI & DATA SCIENCE

01 / ABOUT ME

I like taking messy, real-world data and turning it into systems people can actually trust.

I work where machine learning meets real software: agentic AI, retrieval and forecasting, and the backend and cloud infrastructure that keeps it all running.

I care less about what a model can do in a demo and more about whether it holds up in production: accurate, explainable and safe to rely on.

Education

2023 – 2027
VIT VELLORE
B.Tech · CGPA 9.22
BRANCH
CSE · DATA SCIENCE
Computer Science & Engineering, Data Science specialisation.
OBJECTIVE
SDE & AI ROLES
Seeking high-growth engineering opportunities.

02 / PORTFOLIO WORK

PROJECT CASE STUDIES

[ CLICK TO INSPECT ]

[ PROJECT A ] Agentic AI

Vidura — Agentic HR Assistant

Agentic HR assistant that turns plain-language requests into permission-checked actions for leave, tickets, onboarding and hiring, via a LangChain agent and an MCP server.

Python LangChain FastMCP Pydantic Streamlit
[ PROJECT B ] AIOps · Agents

OpsPilot — AI On-Call Engineer

ReAct agent that detects, localizes and safely fixes broken Kubernetes services. 94% of incidents solved, 69% fewer large-model tokens, 0 unsafe commands run.

Python Kubernetes LLM Cascade
[ PROJECT C ] Hybrid RAG

Satyori — Hybrid RAG Lab

Hybrid RAG for property due diligence: BM25, dense and SPLADE retrieval fused and reranked. 0.744 nDCG@10 on BEIR SciFact, +11% over BM25, with cited answers and conflict detection.

Python SPLADE Cross-encoder BEIR Streamlit
[ PROJECT D ] Full Stack

Fullstack POS System

Full-stack point-of-sale application for inventory management, billing, sales tracking, and business analytics.

React Spring Boot MySQL

03 / SKILLS MATRIX

TECHNICAL SKILLS

[ 6 DISCIPLINES ]

AI & LLM Systems

01

Agents, retrieval and guardrails for LLM apps that act on real systems.

  • Agentic AI
  • Hybrid RAG (dense + BM25 + SPLADE)
  • LangChain
  • MCP / FastMCP
  • Pydantic
  • ChromaDB
  • Cross-encoder reranking
  • Retrieval evaluation (BEIR, nDCG)
  • LLM routing & cascades
  • Guardrails
  • Prompt engineering
  • Groq

Used inOpsPilot · Vidura · Satyori

Machine Learning

02

Forecasting, gradient boosting and vision models, measured against baselines.

  • PyTorch
  • Scikit-learn
  • CatBoost
  • YOLOv8 / OpenCV
  • Regression
  • Forecasting
  • Feature engineering
  • Model evaluation
  • A/B testing
  • GeoPandas

Used inRetail Store Optimization · Speed Bump Detection

DevOps & Cloud

03

Containers, clusters and CI that keep models tested and shipping.

  • Docker
  • Kubernetes
  • kind
  • GitHub Actions
  • CI/CD
  • AWS
  • Oracle Cloud (OCI)
  • Git
  • pytest
  • uv

Used inOpsPilot · AWS and OCI certifications

Backend & Data

04

APIs, services and pipelines that feed models and serve their results.

  • FastAPI
  • REST APIs
  • Spring Boot
  • React
  • MySQL
  • Vector databases
  • ETL pipelines
  • Data cleaning

Used inFullstack POS · Vidura · Satyori

Languages

05

The languages I write every day and the ones I reach for when needed.

  • Python
  • SQL
  • Java
  • C++

Used inAll projects

Analytics & Visualization

06

Turning model output into maps and dashboards people can decide with.

  • Pandas
  • NumPy
  • Matplotlib
  • Seaborn
  • Folium
  • Streamlit

Used inRetail Store Optimization · Satyori

03b / CREDENTIALS

PROFESSIONAL CERTIFICATIONS

[ AWS · ORACLE CLOUD ]

Amazon Web Services
Foundational

AWS Certified Cloud Practitioner

Cloud concepts, security and compliance, core AWS services, pricing and billing.

  • Cloud architecture
  • IAM & security
  • Billing & pricing
Foundational level ID 228594458dfa46e28a7f57813c66cf2b
Oracle University
Professional

OCI Certified Generative AI Professional

LLM fundamentals, RAG with OCI Generative AI, vector search and agent deployment.

  • LLMs
  • RAG
  • Vector search
  • AI agents
Oracle CloudProfessional level
Oracle University
Professional

OCI Certified Data Science Professional

Building, training and deploying models with OCI Data Science and MLOps tooling.

  • Model training
  • Deployment
  • MLOps
Oracle CloudProfessional level
THOUGHTS

05 / CONTACT PIPELINE

LET'S BUILD SOMETHING WORTH REMEMBERING.

Seeking software engineering, backend engineering, and systems-focused developer roles where performant architecture matters.

06 / A PERSONAL NOTE

Greetings from THE ENGINE
ROOM
INDIA · AFTER HOURSNº 026
turn me over ↻
Post Card

Dear reader,

Thank you for wandering through my work.

I love building backend engines, data pipelines and AI systems that quietly hold everything up.

If any of it resonated, I'd love to talk.

— Chunduri Kaushik

₹5 INDIA
PAR AVION · BY AIR MAIL
  • TOYou, the reader
  • Wherever great software
  • gets built
  • PIN404 · not found
Made in IndiaPostage paid · Curiosity Engine

[ Tap the card to turn it over ]