Code

Battle-tested systems, not POCs

Showing 9 of 9 projects

Mossaic

Voice-first research recall built on Moss. Ask your user research out loud, get cited answers in under a second. Dual-surface: public WebRTC voice agent for queries, Slack bot for document ingestion, one shared index.

“Voice-first semantic search, end to end”

Open Source

View Project

AI/ML

distributed-multi-modal-agentic-ai

Showcasing Multi-Model Agents in action on a Distributed Cloud Infrastructure

“Multi-model AI agent orchestration”

Open Source

View Project

AI/ML

perceptra

Conversational AI using GPU Droplets and NVIDIA NIMs

“Production conversational AI”

Open Source

View Project

Cloud

ai-ml-bootstrapper

Quickstart guide incorporating demos/modern best practices for running AI/ML workloads.

“AI/ML infrastructure simplified”

Open Source

View Project

Distributed Systems

pixels-to-cloud

A photo-editor app that aims to inspire developers to create high-performance, scalable, distributed systems.

“High-performance distributed systems”

Open Source

View Project

Cloud

k8s-bootstrapper

Bootstrapping a Production-Ready DigitalOcean Kubernetes Cluster Using Terraform and Argo CD.

“Production Kubernetes, operationalized”

Open Source

View Project

Cloud

k8s-mesh

A pragmatic guide on leveraging the capabilities of Linkerd service mesh on DigitalOcean Kubernetes (DOKS).

“Service mesh architecture”

Open Source

View Project

Cloud

benchmarking-jitsi

WebRTC-based Video Conferencing Platform stress tested on DigitalOcean Droplets.

“Real-time systems, benchmarked”

Open Source

View Project

Cloud

mastodon-on-kubernetes

Setting up Mastodon on DigitalOcean Kubernetes + Managed PostgreSQL + Redis + Spaces.

“Complex deployments on Kubernetes”

Open Source

View Project

About This Portfolio

Common questions from prospects and partners

Are these production systems or proof-of-concept demos?

Production-grade systems, not typical demo apps. Real architecture decisions: event-driven patterns, observability, CI/CD, infrastructure as code. The hard parts that demo apps skip.

What does this portfolio prove?

That I can architect, build, and ship production AI systems. Agent orchestration, event-driven pipelines, real-time voice, distributed systems on Kubernetes. The hard parts demos skip: error handling, observability, deploy paths, failure modes.

Were these built for client engagements or personal projects?

Mix of both. Some were built for clients. Others are systems I run for myself, like the agent sandbox in Argus or the architecture maps in ArchByte. Either way, real architecture, real workloads.