Skip to content
Tanishk Jain
Résumé

Hi, I am

Tanishk Jain

I build the interface and the model underneath it, and I am happiest on the seam where those two have to agree.

I build interfaces forthinking machines.
See the workRésumé
Tanishk Jain, hands joined in greeting

जय जिनेन्द्र

a Jain greeting

Proof of work

Five projects. Each one links to a write-up that states what it does, what it measures, and where it falls short.

All projects
Featured

Parkinson's Voice & XAI

Machine learning2026

Detecting Parkinson's disease from voice samples using explainable machine learning.

Started during a college group project that scored 94.8% accuracy, I suspected our model was cheating. Post-deadline, I launched an independent audit, exposed a speaker-overlap data leak, and rebuilt the entire machine learning pipeline to find the honest clinical truth.

  • Python
  • scikit-learn
  • SHAP
  • LIME
  • pandas
+0.159
Mean ROC AUC optimism from the leak
960
Model runs across both protocols
0.921
Speaker-level sensitivity, honest protocol
Read the case study

Hisaab

Full stack2025

Splitting group expenses to the exact paisa and minimizing required transfers.

Three days after the trip, someone asks who paid for the cab. Hisaab untangles group debt in seconds, splits bills to the exact paisa, and calculates the absolute fewest payments needed to settle up. No awkward follow-ups. Done.

  • Next.js
  • TypeScript
  • Postgres
  • Supabase
  • Row-Level Security
Read the case study

Focus.

Machine learning2026

Real-time webcam attention and gaze tracking using machine learning.

Look away and it notices. A webcam tracker that turns eye contact into an interface. If you lose focus, it responds in real time, proving that machine learning should help you stay present, not just log your distractions after you've already drifted.

  • Python
  • OpenCV
  • MediaPipe
  • NumPy
Read the case study

Fenero Capital

Full stack2026

Institutional financial advisory portal with a zero-hallucination AI assistant.

Securing corporate capital requires trust, not decorative design loops. This is a clean, hyper-focused portal for high-stakes capital advisory, paired with a specialized financial assistant trained to answer firm-specific questions without hallucinating data.

  • React
  • Vite
  • Vercel
  • SEO
Read the case study

Attendance Tracker

Full stack2026

Full-stack workforce attendance management with a spreadsheet-style admin grid.

Managing attendance at scale is a mess of slow grids and broken Excel sheets. This builds a lightning-fast, spreadsheet-style dashboard that untangles employee records and reports in real time, handling the tracking headaches so companies can focus on operations.

  • AngularJS
  • Django REST Framework
  • SQLite
  • Handsontable
  • Highcharts.js
Read the case study

Open in desktop view for the live website preview.

View all projects & archive

Currently building

  • Spot.In progress

    You shouldn't need four different disconnected apps to track your fitness. This gathers your workouts, hydration, sleep, and recovery indicators into a single, cohesive dashboard that finally connects the dots of your training.

    • Next.js
    • TypeScript
    • Postgres
  • guruKoolIn progress

    Stop scrolling through massive PDFs and documentation. Point this tool at any folder of documents, and start asking it questions. You get grounded, accurate answers instantly, without sending your data to external servers.

    • Python
    • RAG
    • Vector search

Experience

Two internships, eighteen months of running a technical club, and one hackathon final. Open a row for the detail.

What I work with

Four domains. Open one to see the tools. Everything listed is something I have actually shipped or trained with.

Try

Alongside
NumPy, pandas, Matplotlib
Tooling
Git, GitHub, Jupyter, Google Colab, Claude Code, Antigravity

About

Not a summary of the projects below. The order it actually happened in, starting from the first program I wrote.

Studying
B.Tech Computer Science, AI and ML
Where
VIT Bhopal, 2023 to 2027
Based in
Noida, India
Recently
ML Research Intern, IIT Delhi
  1. Year one, first semester

    A Tuesday, ten in the morning

    Introduction to Problem Solving: flowcharts, algorithms, pseudocode, then my first real program. By that weekend I had MinGW installed so I could write it again in C++.

  2. That first semester

    Two languages at once

    File handling and the STL on the C++ side, logic building on the Python side. No projects yet, nothing to show anyone. I was still reading the map.

  3. Year two

    The moment code got a face

    Tkinter put a window in front of my code and ODBC put a database behind it. The first time a program of mine was something a person could open, rather than something I ran in a terminal.

  4. Year two, fourth semester

    Structures, and Java

    Data structures and algorithms, then Java, next to the coursework that sits underneath every application: databases, operating systems, networks.

  5. 2024

    The web, before the models got here

    HTML, CSS and JavaScript, learned while AI had barely touched frontend work. The first thing I built was a storefront for a friend, hand-written, no framework holding it up.

    SoleMates
  6. Late 2024

    React, and then a harder question

    A few Frontend Mentor pages, enough to get the shape of React. Then the question no exercise was going to answer for me: what is worth building?

  7. The turn

    Looking for problems, not for prompts

    So I looked outward. For problems still unsolved, and for solved ones explained so badly that the answer never reaches the person who needs it. The second kind is the one I keep coming back to.

  8. October 2024

    Where the record starts

    I did not own a GitHub account until the eighth of October 2024. Everything before it, including a calculator whose one distinction was that it looked good, was never pushed anywhere.

  9. January 2026

    DJUBO, and a stack I had never seen

    First internship. AngularJS 1.x, Handsontable, Moment.js, Highcharts, then Django with Postman for the routes. Hotels and hospitality do not get downtime, so I learned all of it on the job.

  10. June 2026

    IIT Delhi, on long odds

    The résumé got through, then an interview that was machine learning, deep learning and the mathematics under both, harder than I had prepared for. I got in, and the research work started there.

  11. Now

    Placement Season, but I am still building

    Final year at VIT Bhopal. Reading papers I could not have followed a year ago, shipping this site in the evenings, and looking for the next place to be useful.

Certifications

  • ServiceNow Certified System Administrator (CSA)
  • AWS Technical Essentials
  • Applied Machine Learning, University of Michigan
  • LaunchED Data Science
  • Cloud Computing, NPTEL
  • Internet of Things, NPTEL

Find me

Every way in. The line below tells you what each one is for.

Pick one.

GitHubLinkedInInstagramEmailPhoneRésumé
Drop a review