A builder who keeps crossing the line between Android, systems, graphics and AI.
I ship real Android software from a phone, experiment with rendering engines and distributed robotics architecture, and treat AI coding tools as an accelerator — not a replacement for engineering judgment. This site is a map of that work, built directly from my public GitHub record.
One system, four orbits.
Every planet and moon here is a real, publicly visible repository, orbiting a shared core. Drag to rotate, scroll or pinch to zoom, tap a node to open its evidence file. Status color tells you how far along it actually is — nothing here is dressed up as more finished than it is.
3D rendering isn't available on this device — use the project index below instead.
Featured builds
An accessible, non-3D list of the same projects — same data, same links, keyboard navigable.
AppBlur
An idle-touch privacy screen for Android. If you stop touching the screen for a configurable delay, a solid privacy scrim fades in over the display.
AndroidTk
A native terminal / systems-tooling project for Android, iterated through 100+ commits on GitHub, aiming at a real Linux-like on-device workflow.
Phoenix Studio
An in-progress rendering engine built to develop Android games natively, on-device, with its own CI-tracked build pipeline.
AI for All
A content/social app exploring how AI news, tutorials, and research can be made accessible in a feed-style product.
Ros-IM
A designed architecture for a modular, distributed robotics suit system orchestrated with ROS — a documented design, not yet a built system.
Skills, grouped by domain
No percentage meters — just what's core, what's practical, and what's still experimental.
AI-assisted, not AI-authored
Tools like Claude and GitHub Copilot are part of the workflow — for scaffolding, debugging, and speeding up implementation. Architecture decisions, review, and integration are still the engineer's job.
Built from a phone. No computer involved.
AppBlur — the strongest shipped project on this map — I wrote, built, and released entirely from a smartphone: GitHub Actions handles the compile, sign, and release steps that would normally need a desktop and Android Studio.
Smaller, earlier-stage repositories
These are public on GitHub but visibly less mature — most have no README or extensive source inspection available beyond the profile listing. They're included honestly, as experiments, not finished products.
Ai-cloud-phone-
A public experiment exploring a cloud-hosted "phone" concept, visible on GitHub as an HTML-based project.
Ai-cloud-phone-companion
A companion project to Ai-cloud-phone-, also HTML-based and early-stage.
Ai-habit-engine
A public repository exploring an AI-assisted habit-tracking concept.
Project.Vision
An early-stage computer-vision / AI experiment.
Star-data-logger-
An automation-flavored repository, likely a Python + GitHub Actions pipeline for logging GitHub star data.
anti-g
A small public experiment; scope not yet substantiated beyond its presence on the profile.
Source, not claims
Every project card links straight back to its GitHub repository — commit history, workflows, and releases included — so nothing on this page has to be taken on faith.
AppBlur
↗Abhi13-coder/App-blur
AndroidTk
↗Abhi13-coder/Agent-viewer
Phoenix Studio
↗Abhi13-coder/Phoneix-studio
AI for All
↗Abhi13-coder/Ai_for_all
Ros-IM
↗Abhi13-coder/Ros-IM
Ai-cloud-phone-
↗Abhi13-coder/Ai-cloud-phone-
Ai-cloud-phone-companion
↗Abhi13-coder/Ai-cloud-phone-companion
Ai-habit-engine
↗Abhi13-coder/Ai-habit-engine
Project.Vision
↗Abhi13-coder/Project.Vision
Star-data-logger-
↗Abhi13-coder/Star-data-logger-
anti-g
↗Abhi13-coder/anti-g
Abhiram
I'm a B.Tech CSE student and independent builder. I work across Android systems tooling, native rendering engines, distributed-robotics architecture, and practical AI/LLM experimentation — usually starting from constraints (a single phone, no desktop, free CI minutes) rather than ideal conditions. I use AI coding tools deliberately, as an accelerator inside a normal engineering workflow: requirements, architecture, implementation, review, and release.
Right now: pairing AI with my own Python skills — fine-tuning, quantization, and prompt/instruction tuning, run on Colab, Kaggle, and Hugging Face. Also comfortable with HTML, CSS, and JS, plus poster and thumbnail design in the VistaCreate app.