Portfolio

Projects

AI Infrastructure, MLOps, and Platform Engineering projects demonstrating hands-on expertise

Live Projects

AI Infrastructure

AI Research Project Generator

Live

AI-powered research tool with local LLM integration (Ollama, OpenAI), academic search APIs (OpenAlex, Semantic Scholar), and FastAPI REST API for generating comprehensive research projects.

Python
FastAPI
Ollama
LangChain
OpenAlex API
View on GitHub
Platform Engineering

devinpatterson.dev

Live

Personal portfolio website built with Next.js 16, deployed to GitHub Pages via CI/CD pipeline with automated testing, linting, and deployment.

Next.js
TypeScript
Tailwind CSS
GitHub Actions
View on GitHub

In Progress

AI Infrastructure

LLM Serving Benchmark Suite

In Progress

Comprehensive benchmarking framework comparing vLLM, TGI, Ray Serve, and Triton Inference Server for production LLM deployment. Includes batch vs streaming inference comparison, failure isolation testing, latency budgets, and cost per query analysis.

Python
vLLM
TGI
Ray Serve
Triton
Docker
Kubernetes

Completed Projects

AI Infrastructure

Kafka MCP Server

Completed

Custom MCP server enabling AI agents to interact with Kafka streaming infrastructure for real-time operational insights and automated troubleshooting.

Python
MCP
Kafka
FastAPI
AIOps

AI-Powered Root Cause Analysis

Completed

LLM-based workflows for querying servers, analyzing logs, and capturing results to wikis for incident resolution with automated runbook generation.

Claude
MCP
Confluence
Python
Jira

Upcoming Projects

AI Infrastructure
Planned

Production RAG System

Enterprise-grade Retrieval Augmented Generation system with vector databases, evaluation pipelines, chunking strategies, and production deployment pat...

LangChain
Pinecone
ChromaDB
Python
FastAPI
+1 more
MLOps
Planned

MLflow Experiment Tracking Platform

Self-hosted MLflow deployment on Kubernetes with S3 artifact storage, PostgreSQL backend, and integration with model registry for ML lifecycle managem...

MLflow
Kubernetes
Helm
PostgreSQL
S3
+1 more
MLOps
Planned

Kubeflow ML Pipeline

End-to-end ML pipeline on Kubeflow with automated training, hyperparameter tuning (Katib), model serving (KServe), and drift detection using Evidently...

Kubeflow
KServe
Katib
Kubernetes
Python
+1 more
AIOps
Planned

AIOps Observability Platform

Intelligent observability stack with n8n workflow automation, Prometheus/Grafana metrics, Loki logs, and LLM-powered anomaly detection and automated i...

n8n
Prometheus
Grafana
Loki
OpenTelemetry
+2 more
MLOps
Planned

Feature Store with Feast

Production feature store using Feast for ML feature management, online/offline serving, and feature versioning with Redis and BigQuery backends.

Feast
Redis
BigQuery
Python
Kubernetes
Platform Engineering
Planned

K8s Operator for ML Workloads

Custom Kubernetes operator for managing ML workloads with automated deployment, scaling, GPU scheduling, and production-grade reliability features.

Go
Kubernetes
Operator SDK
Prometheus
NVIDIA GPU Operator
Platform Engineering
Planned

Internal Developer Platform (IDP)

Backstage-based developer portal with custom plugins for service catalog, CI/CD integration, infrastructure provisioning, and developer self-service. ...

TypeScript
Go
Backstage
React
Kubernetes
+2 more

Want to see more?

Check out my GitHub for more projects and contributions

View GitHub Profile