I build autonomous agent infrastructure, custom Kubernetes operators, and high-performance backend systems. My work bridges low-level systems engineering with modern AI orchestration—from building lightweight causal neural networks for real-time signal processing to designing tamper-evident governance pipelines for LLM agents.
|
Autonomous agent orchestration platforms, Model Context Protocol (MCP) integrations, OPA policy enforcement, and lightweight neural network architectures for real-time inference. |
Kubernetes controllers and custom CRDs, air-gapped infrastructure monitoring, thread-safe in-memory cache engines, and self-healing systems. |
|
Hybrid retrieval engines combining sparse BM25 encoding with dense vector representations, alongside STFT spectrogram-based speech enhancement models. |
Decoupled web applications utilizing multi-level progressive loading architectures, streaming HTTP APIs, and AI-powered browser developer tools. |
An orchestration platform that runs autonomous AI agents as first-class Kubernetes workloads with Prometheus metrics, Open Policy Agent (OPA) permission checks, and tamper-evident audit trails.
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ LLM Planner ├──────►│ OPA Policy ├──────►│ Prometheus │
│ (Groq/OpenAI) │ │ Permission │ │ Metrics & Audit │
└────────┬────────┘ └────────┬────────┘ └─────────────────┘
│ │
▼ ▼
┌─────────────────────────────────────────────────────────────────────┐
│ Kubernetes Agent CRD Deployment (Isolated Pods) │
└─────────────────────────────────────────────────────────────────────┘
Python · Kubernetes · OPA Policy · Prometheus · Groq / OpenAI API
A 100% offline, Go-based Kubernetes operator designed for air-gapped datacenters that monitors infrastructure telemetry, predicts time-to-impact (TTI) for failures, and executes repair skills via the Model Context Protocol (MCP).
┌────────────────────────┐ ┌─────────────────────────┐ ┌────────────────────────┐
│ Hardware Telemetry Stream ├──►│ Predictive TTI Analyzer ├──►│ MCP Skill Execution │
└────────────────────────┘ └─────────────────────────┘ └────────────────────────┘
Go · Kubernetes Operator · MCP Protocol · React 18 · Telemetry
|
Autonomous AI agent lifecycle governance and operational control platform for Kubernetes.
Tech: |
Predictive telemetry and self-healing infrastructure operator for air-gapped environments.
Tech: |
|
High-performance, thread-safe in-memory key-value cache server with REST API interface.
Tech: |
Lightweight Causal U-Net for real-time speech enhancement on spectrograms.
Tech: |
|
Similarity and Linguistic Filtering system for title verification and duplicate detection. 
Tech: |
Enterprise webmail client featuring a 3-level progressive data loading architecture. 
Tech: |
| Category | Technical Focus |
|---|---|
| Languages | Python · Go · C++ · JavaScript / TypeScript |
| Backend & Systems | FastAPI · Node.js / Express · REST APIs · gRPC / Protobuf · Multithreading |
| AI & Machine Learning | PyTorch · LLMs · RAG · Model Context Protocol (MCP) · LangChain · STFT Processing |
| Cloud & Infrastructure | Kubernetes (Controllers, CRDs) · Docker · OPA (Open Policy Agent) · Prometheus · Linux |
| Databases & Vector Stores | Pinecone · Redis · PostgreSQL · In-Memory Caches |
|
|
|
- Agentic Governance & Safety: Building execution sandboxes, policy engines, and audit trails for LLM agents.
- Kubernetes Controllers & Operators: Extending K8s APIs to automate complex infrastructure maintenance and monitoring.
- Low-Latency AI & Audio Systems: Designing memory-efficient neural networks for real-time signal processing on edge devices.
- Hybrid Retrieval & Search Architecture: Combining sparse lexical search (BM25) with dense vector embeddings for precise domain retrieval.