ML Pipeline

Spring Boot

R²=0.96

PostgreSQL

TF-IDF

// Software Engineer — Backend & AI/ML

Shivansh Thakur.

View My Work → Get In Touch ⬇ Resume

Scroll

// 01 — About

Not just a dev.
An engineer.

Hey — I'm Shivansh. I build systems that scale, models that learn, and backends that don't break at 3am. Based in India, thinking globally.

I'm obsessed with the intersection of backend engineering and AI/ML — the place where distributed systems meet intelligent computation. I don't just ship features, I architect solutions.

Whether it's designing an RLHF-style evaluation pipeline, hitting R²=0.96 on response quality scoring, or building interval-tree algorithms for real-time scheduling — I go deep.

Currently pursuing B.E. CS&E at Chandigarh University, but honestly, I learn most from shipping real projects and reading research papers at 2am.

% Accuracy (IEEE Paper)

★ CodeChef Rating

Production Projects

IEEE Publications

// shivansh.config.ts

const shivansh = {

  role: "Backend + AI/ML Engineer",

  location: "India 🇮🇳",

  status: "Open to SDE Roles",

  stack: [

    "Java", "Python",

    "Spring Boot", "FastAPI",

    "ML Pipelines", "NLP",

],

  superpower: "System Design",

  currentlyBuilding: "AI eval systems",

  leetcode: "Knight 🏅",

  codechef: 5 + "★",

// 02 — Skills

The toolkit.

⚡

Languages

Java Python C C++

🔧

Backend

Spring Boot FastAPI REST APIs Microservices System Design

🧠

AI / ML

Machine Learning NLP TF-IDF Gradient Boosting Transfer Learning RLHF

🗄️

Databases

PostgreSQL MongoDB SQL Data Pipelines

☁️

Cloud & DevOps

AWS EC2 AWS S3 Docker Git CI/CD

📐

CS Fundamentals

DSA OOP OS Networks DBMS

// 03 — Projects

Things I built.

🤖

Scorix AI

R² = 0.96 response quality scoring

LLM Evaluation, Ranking & Feedback System. Built an ML-based evaluation engine using TF-IDF, cosine similarity, and Gradient Boosting. Features RLHF-style continuous retraining from human feedback.

FastAPI TF-IDF Gradient Boosting RLHF Python

📊

AIBench

Multi-metric LLM benchmarking platform

Comprehensive LLM Benchmarking & Evaluation Platform evaluating responses across semantic similarity, relevance, and coherence. Real-time analytics dashboard with database-backed data pipelines.

Python NLP PostgreSQL Data Pipelines Analytics

📅

↗

Slotify

O(log n) interval-tree conflict elimination

Smart Scheduling & Booking System with real-time availability management. Architected a full-stack platform with dual PostgreSQL/MongoDB database layer and conflict-free booking logic.

Spring Boot REST APIs PostgreSQL MongoDB Vercel

🖼️

Serverless Image Resizer

Event-driven · CDN delivery via CloudFront

Serverless image processing system on AWS — dynamic resizing, format conversion, and CDN delivery. Pre-signed uploads + event-driven architecture using S3 triggers and Lambda functions.

AWS Lambda S3 CloudFront Serverless Python

📈

Stock Market Volatility Simulator

Quantitative risk classification engine

Quantitative finance project that analyzes stock volatility, computes statistical risk metrics (VaR, Sharpe ratio, std dev), and classifies investment risk using Python-based ML models.

Python Quantitative Finance ML Risk Metrics Statistics

🔍

TruthLens

99.04% classification accuracy

Fake News Detection system — the production implementation behind the IEEE-published research. Uses RNNs with NLP preprocessing to classify news articles at scale, outperforming SVM and Naive Bayes baselines.

Python RNN NLP TF-IDF Deep Learning

// 04 — Research

Published work.

IEEE

ICCCNT
IIT Indore

Fake News Detection with Recurrent Neural Networks

99.04% accuracy · 44,898 articles · Outperforms SVM & Naive Bayes

IEEE

ICCCNT
IIT Indore

Plant Disease Classification Using Transfer Learning with MobileNet

95% validation accuracy · MobileNet fine-tuning · Multi-class pathology

// 05 — Experience

Where I've trained.

Jun 2025 – Jul 2025

Software Engineering Trainee

Intel Corporation (via Chandigarh University)

Developed end-to-end ML pipelines covering dataset preprocessing, feature engineering, and model evaluation under Intel's engineering framework
Built REST APIs for real-time ML inference workflows, integrating model outputs into production-ready endpoints

Aug 2023 – Present

B.E. Computer Science & Engineering

Chandigarh University · CGPA 7.67/10

Secretary, Blockchain Club — organized department-wide seminars managing 100+ student community
Focus: DSA, OS, DBMS, Computer Networks, OOP

// 06 — Achievements

The numbers.

⚔️

Knight

LeetCode Rating

Top competitive tier

⭐

5★

CodeChef

Elite competitive programmer

🎯

% IEEE Accuracy

Fake News Detection model

📈

0.96

R² Score

Scorix AI evaluation engine

🏛️

IEEE Published

ICCCNT at IIT Indore

👥

100+

Community Led

Blockchain Club Secretary

// 07 — Contact

Let's talk.

Got an opportunity?

I'm actively looking for SDE roles in backend engineering and AI/ML. If you're building something cool and want someone who actually cares about the craft — let's talk.

thakur.shivansh.pundir@gmail.com

✓

Message sent! I'll get back to you soon.