- Basekick-Labs/arc#769 — ✅ Merged ·
ingest/msgpack: log numeric host coercion at debug level (Go, performance-verified, 3 review rounds) - firstcontributions/first-contributions#124856 — ✅ Merged · Added contributor entry
AI INFRASTRUCTURE · BACKEND · SYSTEMS
I build the infrastructure that keeps AI systems running when things go wrong.
LLM providers fail. Requests get rate-limited. Latency spikes. Models become unavailable.
That's the layer I'm interested in.
I build systems around AI that handle failure instead of pretending it doesn't exist:
- ⚡ Intelligent gateways and provider routing
- 🛡️ Automatic failover and resilience
- 🚦 Rate limiting, retries and circuit breakers
- 📦 Response caching and streaming
- 📊 Observability and infrastructure metrics
- 🤖 Experiments with agents and local AI
- 🧠 Low-level machine learning from scratch
If it breaks, I build a layer around it.
Currently focused on AI Infrastructure, LLM Routing, Resilience Engineering, and Agent Systems.
Bug manufacturer 🐛 · Professional bug exterminator 🔨 · Occasional Rust compiler argument loser 🦀
Languages
Go · Rust · Python · C · JavaScript
AI / LLM
LLM Gateways · AI Agents · Embeddings · Semantic Search · LLM Streaming
Infrastructure
Docker · Linux · Prometheus · OpenTelemetry · Redis · REST APIs
Exploring
Distributed Systems · Provider Routing · Circuit Breakers · Observability · Local AI
AI Gateway · Failover · Resilience
An intelligent proxy designed to sit between AI applications and model providers.
It handles provider failover, smart routing, rate limiting, caching and observability.
Stack: Go HTTP Prometheus YAML OpenAI-compatible APIs
- 🔄 Provider failover
- 🎯 Smart routing
- 🚦 Rate limiting
- 💾 Response caching
- 📡 Prometheus metrics
- 📊 Live dashboard
- 🔁 Retry policies
- 🏥 Health checks
Neural Networks · From Scratch · Android
An experiment focused on understanding AI from the inside instead of hiding everything behind frameworks.
Built from an Android phone using Python + NumPy + Termux.
No PyTorch. No TensorFlow. No shortcuts.
Stack: Python NumPy Android Termux
- 🧮 Neural networks from scratch
- 👁️ Vision experiments
- 🧠 Memory systems
- 🎙️ Voice experiments
- 🤖 Agent experiments
- ⚙️ Optimization and training
Semantic Code Search · Local AI Embeddings
A code search engine focused on meaning instead of keywords.
Search for what the code does, not just what the code is called.
Stack: Rust AI Embeddings Semantic Search
Experimental Python Project
A project currently under development.
Status: STEALTH
Low-Level Systems · C
A low-level C project focused on keeping things simple, raw and fast.
Because sometimes Go is just too high-level.
Stack: C Systems Programming Low Level
Rust · Systems Engineering
A Rust systems project under Axiom Systems Tech.
Stack: Rust Systems Engineering
| Project | Focus | Status |
|---|---|---|
| 🛡️ SentinelFlow | AI gateways · failover · routing | 🟢 Active |
| 🧠 Cerebro Zero | Neural networks · agents · AI from scratch | 🟡 Experimental |
| 🔍 Semcode Search | Semantic code intelligence | 🟡 Building |
| 🌩️ Nuvora | Python project | 🔒 Stealth |
Go · Rust · Python · C
AI Infrastructure
│
├── LLM Gateways
├── Provider Failover
├── Smart Routing
├── Rate Limiting
├── Caching
├── Observability
│
└── Agent Systems
├── Tool Use
├── Memory
├── Reasoning
└── Reliability
"Make it work, make it right, make it fast — in that order."
I'm open to collaborations around:
AI infrastructure · LLM tooling · backend systems · resilience engineering · low-level experiments
If you're building something that needs to survive production — or something that simply refuses to work at 3am — I'm probably interested.
