I am a Simulation Architect in progress, building systems that turn abstract math into runnable, inspectable, and teachable software.
My working style is proof-driven:
- Understand the concept deeply.
- Build a numerical engine from first principles.
- Validate behavior with tests and checks.
- Render the concept visually for clarity.
This is my main body of work and the best entry point for recruiters.
Open Master Showcase Repository
| Project | Core Idea | Engineering Outcome |
|---|---|---|
| LA-Visual-Engine | 2D transformations and determinant intuition | Matrix operations proved via geometric motion |
| LA-Visual-Engine-3D | 3x3 transforms in full space | Lattice deformation and basis-vector interpretation |
| Basis-Translator-3D | Change of basis | Coordinate-frame translation with fixed physical anchor |
| Cramers-Rule-3D | Linear systems as geometry | Determinant-ratio solving visualized in 3D |
| Eigenvector-Explorer-3D | Stable directions and eigenstructure | Eigenvectors and eigenvalues made spatially explicit |
An end-to-end AI game engineering project spanning RL environment design, real-time streamed inference UX, and production deployment hardening.
Open Master Repository (Full Journey + Architecture)
| Layer | Repository | What It Demonstrates |
|---|---|---|
| Part 1 | anagram-quest-openenv | RL-compatible environment design, reward shaping, API modeling |
| Part 2A | anagram-quest-server | FastAPI game orchestration, SSE streaming, MDP banking logic, production stability fixes |
| Part 2B | anagram-quest-frontend | Premium Next.js UX, Player-vs-AI mode, reconnect-safe real-time game experience |
| Part 2C | anagram-quest-qwen3-0.6b-grpo-mlx | Model release packaging, model card, and versioned weight publishing workflow |
Live demos:
Cramer's Rule 3D A geometric solver where linear systems become visible 3D volume transformations. |
Eigenvector Explorer 3D A spatial engine showing why certain directions remain invariant under transformation. |
- I do not stop at notebooks: I ship end-to-end systems.
- I combine mathematical depth with software execution.
- I can explain hard concepts visually for non-specialists.
- I build with reproducibility and deployment in mind.
- Physics simulation systems (starting with N-body dynamics)
- RL environments and agent workflows
- Technical AI systems for simulation pipelines
- OpenUSD / Omniverse-aligned long-term path
I am actively open to internships, research collaborations, and builder roles in:
- Simulation and scientific computing
- Technical Python and backend systems
- RL environments and agent workflows
- Visualization and graphics-driven education tools
Kanpur, India • Building globally relevant simulation systems