Systems Engineer • Rust Developer • Local AI Builder • Founder of AarambhDevHub
I build high-performance systems, Rust-native AI experiments, developer tools, and infrastructure projects from first principles.
pub struct DarshanVichhi {
pub alias: &'static str,
pub role: &'static str,
pub org: &'static str,
pub focus: [&'static str; 6],
}
impl DarshanVichhi {
pub fn new() -> Self {
Self {
alias: "Aarambh",
role: "Systems Engineer • Rust Developer • Local AI Builder",
org: "AarambhDevHub",
focus: [
"Rust-native AI systems",
"high-performance backend infrastructure",
"distributed systems from scratch",
"developer tools and proc macros",
"local-first neural memory experiments",
"teaching real engineering by building",
],
}
}
}I am Darshan Vichhi, also known as Aarambh. I work on systems programming, backend infrastructure, Rust frameworks, developer tooling, and local-first AI systems.
My goal is simple: build real systems, understand every layer, and teach the engineering behind them.
I focus on:
- Rust-native AI and local memory systems
- high-performance web frameworks
- distributed systems built from scratch
- developer tools and code generation
- infrastructure, networking, and security tools
- educational engineering projects for AarambhDevHub
| Area | What I am building |
|---|---|
| Rust AI Systems | Local-first AI, LLM architecture, neural memory, inference, quantization, and self-learning experiments |
| Systems Engineering | Databases, message queues, proxies, container runtimes, networking tools, and distributed systems |
| Frameworks | Rust and Go web frameworks focused on performance, routing, middleware, and developer experience |
| Education | Building projects from scratch and explaining the design decisions through AarambhDevHub |
| Project | Description | Focus |
|---|---|---|
Aarambh AI |
Rust-native LLM research project | tokenizer, transformer, training, inference, quantization, LoRA/QLoRA, GRPO, self-learning |
Manas |
Local-first AI memory system in Rust | neural memory, self-growing learning, persistent local knowledge, CPU-first experimentation |
mini-tensorflow |
Educational deep learning framework | tensors, autograd, neural networks, SIMD optimization |
APEX-1 |
Educational LLM architecture project | MLA, MoE, GRPO-style reasoning concepts |
| Project | Language | Description |
|---|---|---|
Ajaya |
Rust | High-performance Rust web framework |
Rudra |
Go | Fast Go web framework |
Ignitia |
Rust | Rust web framework with routing and middleware |
Blaze |
Go | Go web framework with HTTP/2, WebSocket, caching, validation, and compression |
Vega |
Rust | Next.js-inspired Rust framework with compile-time file-based routing |
Capsules |
Rust | Backend for a digital time capsule platform |
| Project | Description |
|---|---|
typebridge |
Generate TypeScript, Python, Go, Swift, Kotlin, Zod, GraphQL SDL, and JSON Schema from Rust types |
Animato |
Renderer-agnostic Rust animation library |
Spanda |
Animation system for WASM, TUI, Bevy, and native rendering |
Scenix |
Rust-native 3D scene toolkit for native and WASM apps |
dev-proxy |
HTTP recording, replay, mocking, and debugging proxy |
| Project | Description |
|---|---|
mini-database |
Graph database with a query engine |
mini-kafka |
Distributed message queue |
mini-redis |
In-memory key-value store |
mini-docker-rust |
Container runtime using namespaces and cgroups |
mini-git |
Version control system from scratch |
mini-p2p |
Peer-to-peer file sharing network |
query-engine |
Distributed SQL query engine |
WsForge |
WebSocket load-testing tool |
| Project | Description |
|---|---|
pingora-waf |
Web application firewall inspired by Pingora |
exam-cheating-detection |
AI proctoring and cheating detection system |
multi-cam-face-tracker |
Multi-camera face tracking and alerting |
Gupti |
One-time encrypted secret sharing app |
This section avoids third-party stat cards so it stays stable in GitHub preview.
| Repository | Why to check it |
|---|---|
AarambhDevHub/aarambh-ai |
Rust-native LLM architecture and training blueprint |
AarambhDevHub/manas |
Local-first AI memory and self-growing learning experiment |
AarambhDevHub/ajaya |
Rust web framework work |
AarambhDevHub/rudra |
Go web framework work |
aarambh-darshan/typebridge |
Rust proc-macro based type generation |
AarambhDevHub/pingora-waf |
Security and networking infrastructure |
01. Ship source, not hype.
02. Benchmark before making performance claims.
03. Prefer explicit architecture over hidden magic.
04. Build small working systems before large ones.
05. Keep learning loops measurable.
06. Teach internals, not just APIs.
07. Stay local-first when it makes sense.
08. Optimize only after understanding the bottleneck.
A practical guide for students and developers who want to stop watching tutorials and start shipping real projects.
Start Building. Keep Building.
© 2026 Darshan Vichhi — Aarambh