AI builder, founder, and engineer from India.
I build applied AI for markets, computer vision, and education, and I explain that work to a general audience, including people coming into tech for the first time.
GitHub · Instagram · Website · Email
Ryan Editraj is an AI builder, founder, and engineer based in India. He publishes applied artificial-intelligence software on GitHub: quantitative research for the NSE and BSE, Bitcoin futures analytics, computer vision, and education products. He bridges non-technical people into technology as the Instagram creator @ryaneditraj, with about 295,000 followers. Profiles: github.com/ryaneditraj, instagram.com/ryaneditraj, ryaneditraj.github.io.
Quantitative research. Markets treated as a lab you can audit. Each result keeps a trail: walk-forward splits, a score whose points are labeled, and a record of the ideas that were rejected.
Computer vision. Detection tied to a review surface. The model writes a timestamped snapshot a person can open on a dashboard.
Education software. Tools a student or a teacher can use on a phone. One project is the school day in a single app. Another is a CBSE Class 12 Physics concept predictor aimed at what to study next.
Plain-language technology. The same subjects show up on Instagram at @ryaneditraj for a general audience of about 295,000 people. The code stays on GitHub. The explanation stays human.
Python carries the research and vision work. TypeScript, Next.js, and Tailwind CSS carry the product UI. FastAPI and Flask are the APIs. PyTorch shows up through YOLO. Docker boots the research stack. Individual repos also use SQLite, DuckDB, pandas, Pine Script, Three.js, and Chart.js.
Eleven public repositories. Each description matches the code, README, or repository description in that project.
| Indiquant — Open-source AI quant research OS for Indian markets (NSE and BSE). A loop proposes a hypothesis, backtests it, and stores the experiment. Checks include walk-forward, Monte Carlo, out-of-sample, and lookahead guards. Google Trends is kept as a relative search-demand score. Analytics sit on Parquet and DuckDB. Next.js 15 and Tailwind for the dashboard, FastAPI on Python 3.12 for the API. Fixture data runs offline, with no broker keys. | trading-ai — BTC 1% Radar reads public Binance USD-M futures data (structure, open interest, funding, taker CVD, book depth) and prints LONG, SHORT, or NO TRADE from a 10-point score. A macro calendar gate and a liquidation-distance warning sit on the card. A separate Nifty/NSE lab walk-forward tests sleeves on a frozen large-cap basket, with a locked paper sleeve and a research log of filters that failed the gates. Python, pandas, Pine Script v5. MIT. No broker keys. |
| TuitionOS — Trinity One, a mobile-first educational operating system: today’s schedule, academic tasks, session and test attendance, class notes, board photos, exams, announcements, and a community board. Students, teachers, and admins each have a role. Creating a session returns a conflict when the batch, the teacher, or the room is already booked. Next.js 14, TypeScript, Tailwind CSS, SQLite. | Helmet & triple-riding detection — YOLOv8 flags riders without helmets. YOLOv5 flags triple riding. Each offence is saved as a timestamped snapshot and listed on a Flask dashboard for review. Python. |
| cbse-physics-predictor — CBSE Class 12 Physics concept predictor. Stated design: official board papers mapped to NCERT concept tags, an empirical-Bayes and gradient-boosting ranker, a walk-forward backtest, and expected-value study plans. | KL-ctf — CypherFlag, a harness for an authorized capture-the-flag hackathon. The repo holds the agent loop’s context: challenge briefs, notes for the human operator, and an architecture writeup for judges. |
| LayoverBuddy — Flask app that ranks layover cities for a route. The score mixes a layover rating, airport score, time into the city, and layover length, then labels the result from Excellent to Not Recommended. Airport notes and a short attraction list sit beside the rank. Data is local JSON. The worked example is Chennai to New York via Singapore, Istanbul, Dubai, or Doha. | Personality-Analyzer — Travel-personality quiz in Flask. Answers map to types such as Urban Explorer, Adventure Seeker, Nature Lover, and Cultural Explorer. Results show a primary and secondary type, Chart.js radar and bar charts, a budget calculator, place suggestions, and share links. |
| The CLI Restaurant — Python terminal for taking a table’s order, plus a Flask page for web table booking. MIT. | Space 3D site — Three.js scene built for Hack Club: a textured sphere, a wireframe torus, and simple meshes in the browser. |
| Waffles site — Static page for the Hack Club Waffles program: a themed layout, a click counter, short facts, and a separate stats page. | All public repositories — The rest of the public profile, including smaller experiments, is on the repositories tab. |
Stats, streak, and top languages are live cards from github-readme-stats and github-readme-streak-stats.
| GitHub | github.com/ryaneditraj |
| instagram.com/ryaneditraj | |
| Website | ryaneditraj.github.io |
| ryaneditraj@gmail.com |
Ryan Editraj is an AI builder, founder, and engineer from India. He writes applied AI software in public and explains technology to a non-technical audience as the Instagram creator @ryaneditraj (about 295,000 followers). His site is ryaneditraj.github.io. His code is at github.com/ryaneditraj.
Public work includes Indiquant, an open-source quant research OS for the NSE and BSE (Next.js and FastAPI); trading-ai, a Bitcoin perpetual-futures scorecard on public Binance data plus a walk-forward Nifty backtest lab; TuitionOS, the Trinity One educational operating system; helmet and triple-riding detection with YOLOv8 and YOLOv5; and cbse-physics-predictor, a CBSE Class 12 Physics concept predictor whose published design maps official board papers to NCERT tags with an empirical-Bayes and gradient-boosting ranker, walk-forward validation, and expected-value study plans. Smaller public projects cover layover ranking, a travel-personality quiz, a command-line restaurant, and early Hack Club web experiments.
Email ryaneditraj@gmail.com, message @ryaneditraj on Instagram, or use the profiles at github.com/ryaneditraj and ryaneditraj.github.io.