Ritik Roushan RanaNew Delhi

Ritik Roushan Rana

Machine Learning Engineer — focused on deep RL, computer vision and systems that hold up under pressure.

01 — About

Ilikesolvingproblemswithdata,andbuildingthemodelsthatdoit.

Final-year CSE (InfoSec) student at VIT Vellore, focused on machine learning, data science and AI, with cybersecurity on the side. Solid grounding in programming and DSA; I'd rather reach for a data-driven answer than a guess.

Right now I'm building real experience in predictive modelling and data analysis, and going deeper into neural networks and reinforcement learning. I also build mobile apps in Flutter.

Curious and quick to adapt. Looking to build things that matter and grow in ML, data science and AI.

ritikrana8596@gmail.com
Ritik Roushan Rana
New Delhi · 2026

02 — Selected work

Things I'vebuilt

Four projects, from reinforcement learning to supply-chain defence

Optifolio — Overview
Optifolio — Holdings
Optifolio — RL rebalance
Optifolio — Stock detail
Optifolio — Analytics
Optifolio — News & alerts
Optifolio — AI assistant
Overview
01/04

Optifolio

Fintech · Reinforcement learning · Solo build

A Deep Q-Learning portfolio rebalancer that learns its policy from historical market data, wrapped in a Flutter app for goal setting, backtesting and rebalancing suggestions.

Domain
Fintech
Core
Deep Q-Learning
Surface
Flutter iOS
MarketMind — Dashboard
02/04

MarketMind

Fintech · Explainable ML · Solo build

Next-day directional prediction for 15 large-cap equities. XGBoost over price and sentiment features, FinBERT scoring the day's headlines, and per-feature TreeSHAP attributions behind every call — with a plain-English rationale written only from those attributions, so it cannot invent a story the model did not tell.

Domain
Fintech
Core
XGBoost + FinBERT
Surface
FastAPI + React
VTour — Campus guide
VTour — Guided tour
VTour — Campus map
VTour — Location
Campus guide
03/04

VTour

Campus · Mobile product · UI/UX + build

A virtual campus tour for new students — maps, departments and facilities in one Flutter app, designed around a single uninterrupted navigation flow.

Domain
EdTech
Core
Guided tours
Surface
Flutter iOS
04/04

SCBF

Security · Temporal graph networks · Patent pending

Supply Chain Behavioral Fingerprinting: eBPF records what a package actually does while it installs, a Temporal Graph Network turns that into a behavioural fingerprint, and it is scored against what legitimate packages of its type do. Matches the OSCAR benchmark's F1 at ~30x lower per-package latency.

Test F1
92.3%
ROC-AUC
0.979
Per package
~3s

03 — Path

Where I'vebeen

Internship·Martvalley Online Pvt. Ltd

Artificial Intelligence Intern

May 2025 — Jul 2025

  • Shipped real-time AI modules across 5+ development projects, cutting end-to-end processing latency by 30% and lifting system throughput by 25%.
  • Took models to production: optimised the ML training workflows and integrated the resulting models into live systems.
  • Worked the image pipeline: collaborated on pipelines handling 5,000+ images a week, raising model accuracy by 5%.

04 — Toolkit

What I workwith

Click a cap, or type its corner letter. Space clears the display.

05 — Contact

Got something worth building?