Uses
A curated list of tools, technologies, and hardware I use for AI engineering, distributed systems development, and high-performance backend applications.
AI & Data Engineering
- PyTorch with Lightning for rapid model development. Particularly useful for quick prototyping and production-ready AI systems.
- For data warehousing, I work with Amazon Redshift and Snowflake for large-scale analytics.Apache Spark for distributed data processing.
- Subscription systems built with Stripe for billing, Apache Kafka for event streaming, and custom Go microservices for high-performance processing.
- Databricks for unified analytics, dbt for data transformations, and dbt Semantic Layer for metrics.
Development
- VS Code with Windsurf for AI development. Essential extensions: GitHub Copilot, Python, Go, Docker, and Thunder Client for API testing.
- GoLand for complex Go projects, especially when working with large-scale microservices and performance-critical code.
- High-performance web services in Go using Gin and Echo. Custom middleware for rate limiting, caching, and observability.
- Database optimization with PostgreSQL (partitioning, indexing strategies), MongoDB for document storage, and Redis for caching and real-time features.
- Cloud-native development with Kubernetes, Docker for containerization, and Terraform for infrastructure management.
- Observability stack: OpenTelemetry for instrumentation, Grafana for visualization, and Prometheus for metrics collection.
LLM & RAG Development
- LangChain for RAG pipelines with custom Go services for high-performance retrieval.Cohere Go SDK for embedding generation.
- Vector stores: Pinecone for managed deployments, Qdrant for self-hosted with high QPS requirements, and ChromaDB for rapid prototyping.
- OpenAI and Anthropic SDKs for LLM integration. Custom caching and rate limiting middleware in Go.
- Document processing with Unstructured for parsing, custom Go services for high-throughput preprocessing, and Redis for caching embeddings.
- Monitoring with LangSmith for tracing, custom Prometheus metrics for performance, and RAGAS for RAG evaluation.
Workspace
| Workstation | Custom built AMD Threadripper, 128GB RAM, RTX 4090 - Perfect for training and fine-tuning models |
|---|---|
| Operating System | Linux (Pop!_OS) - Perfect balance of stability and customization |
| Terminal | Alacritty with tmux and Oh My Zsh |
| Monitors | Dual 4K 32" Dell U3219Q + LG 34" Ultrawide |
| Keyboard | Keychron K8 Pro (Red switch) |
| Mouse | Logitech MX Master 3 |
| Laptop | Macbook Pro M4 14" |
| Cloud Platforms | AWS (Primary), GCP for AI/ML workloads |
| Headphones | Audio Technica ATH-M50x/Apple Airpods |
| Microphone | Blue Yeti |