Kunal Vats
Building intelligent systems that actually ship.
Computer Science undergraduate specializing in Artificial Intelligence & Data Science — turning trained models into REST APIs and containerized workflows that run inside real applications.
About
A Computer Science undergraduate who treats machine learning as an engineering discipline, not a one-off experiment.
Foundations first
My foundation combines core computer science fundamentals — data structures, algorithms and system design — with applied machine learning. I build primarily in Python, SQL and C++, which keeps me equally comfortable reasoning about a model and reasoning about the system it has to run inside.
Applied machine learning
Day to day, that means using tools like PyTorch, Scikit-learn and Pandas to clean complex datasets, engineer features, and train predictive models — the unglamorous work that decides whether a model is actually usable.
Built for production, not just accuracy
What defines my approach is a focus on production utility. Across my projects in NLP and predictive analytics, I don’t stop at model accuracy on a local machine — I focus on packaging models into clean REST APIs and containerized workflows so they can actually run inside functional applications.
I’m looking to bring this combination of engineering discipline and practical AI problem-solving to an innovative team, where I can build reliable, production-grade software and make an immediate impact.
Technical skills
Organized by where each tool sits in the path from raw data to running software.
Programming
AI & Machine Learning
Frameworks & Libraries
Engineering
Featured projects
Portfolio builds exploring NLP, predictive analytics and production ML — each one structured to move from a trained model into something an application can call.
From model to production
A trained model that never leaves a notebook isn’t software. Every project here is built around the same underlying path, from raw data to something an application can actually call.
-
01
Data
Start from raw, often messy source data — the shape it actually arrives in.
-
02
Processing
Clean, validate and structure it with Pandas and SQL before anything gets trained.
-
03
Feature engineering
Turn raw fields into signals a model can actually learn from.
-
04
Model
Train and evaluate with PyTorch or Scikit-learn, judged on more than one metric.
-
05
API
Wrap the model behind a clean REST interface, not a script that runs once.
-
06
Containerized workflow
Package it so it deploys the same way every time, on any machine.
-
07
Application
Ship something a real application, and a real user, can call.
Learning journey
No shortcuts taken — the same three stages show up again in every project on this site.
Fundamentals, deliberately
Data structures, algorithms and system design first — the parts of computer science that don’t go stale when a new framework does.
Applied AI & data science
Moving from theory into practice with PyTorch, Scikit-learn and Pandas — cleaning real datasets and training models against real, imperfect data.
Thinking in production
Learning that a model is only half the job — the other half is REST APIs and containerized workflows that let it run inside something people actually use.
Where the code lives
Every project on this page is backed by real code. No dashboards, no fabricated metrics — just the repositories.
Let’s build something
Open to internships, entry-level roles and collaborations where AI has to survive contact with a real application.
- Email hello@example.com
- Phone +91 86038 60535
- LinkedIn linkedin.com/in/imkunalvats
- GitHub github.com/imkunalvats