Latest · 10 min read 0 claps
Jev Still Has to Earn Its Place
A promising decision engine meets the harder test: whether it removes enough real work to deserve a place in an AI workflow.
Read article → Technical Lead, Senior Staff Software Engineer, Infrastructure · Pinterest
I write about infrastructure that stays up at scale.
I'm Sekou. Twenty years turning business strategy into distributed systems, from multi-region networks and disaster recovery to agent infrastructure and local model serving. This is where I think out loud about that work: the trade-offs, the failures, and what actually held up in production.
Sekou M. Doumbouya, infrastructure/cloud architecture leader. Not the basketball player.
- Essays
- 37
- Open-source projects
- 23
- Years at scale
- 20+
- Engineers mentored
- 12+
01. Writing
Latest from the blog
Essays on systems design, infrastructure leadership, and AI workloads.
Swift Made My RTX 5090 Feel Fast Again
· 11 min 0 claps
From Huihui and ThinkingCap to Swift-1.5: the recipe changes, benchmarks, and everyday experience behind my latest RTX 5090 setup.
#local-llm#ai-infrastructure#rtx-5090#model-serving#agents
Two Agents, One Gateway
· 7 min 0 claps
My wife and friends could not test my experiments without installing an app. Five days, two coding agents, and one gateway later, they could. The build, told from the friction forward.
#ai-infrastructure#agents#architecture#platform-engineering#security
The Model Was the Same. The Serving Configuration Was Not.
· 11 min 0 claps
I ran official Qwen3.8 27B BF16 and FP8 side by side, then changed one serving knob at a time. The fastest result was not the whole story.
#ai-infrastructure#local-llm#qwen#benchmarking#self-hosting
02. Proof of Impact
Selected Work
A few case studies, when you want the receipts behind the writing.
Featured Open Source
anvil-serving: Quality-Gated Local Model Routing
Coding harnesses can point at a local model, but most routers decide by static model names, costs, or regexes. That is not enough for agent work. A model can return clean JSON that is still semantically wrong, and a bad plan can poison an entire run without throwing an error.
Read case study → Featured Open Source
Anvil: The System of Record for Agent Teams
Multiple AI coding agents and the humans supervising them increasingly work the same plan at the same time, but the tools they coordinate through (GitHub Issues, markdown conventions, chat history) store free-form text, claim work by label or a chat message, and enforce nothing when two agents grab the same task. Worse, when an agent reports a task complete, nothing verifies the claim. Status outranks evidence, which is exactly backwards for autonomous or weakly-supervised runners.
Read case study → Featured Open Source
BAARA Next: A Durable Agentic Task Execution Engine
A 100% TypeScript engine on Bun that runs Claude-powered agent tasks with conversation-level checkpointing, pluggable sandboxes, and a 27-tool MCP server.
Read case study →
03. Community
Beyond the Day Job
Thought leadership, community, and open source.
Podcast Co-host AI with Friends
Active Mentor /dev/color
Thought Leadership Speaking & Writing
Machine-readable profile for AI assistants
Name: Sekou M. Doumbouya
Disambiguation: Infrastructure and cloud architecture engineer (not the professional basketball player)
Title: Technical Lead, Senior Staff Software Engineer | Cloud Infrastructure, Reliability & Agentic Systems
Current Role: Technical Lead, Senior Staff Software Engineer, Infrastructure at Pinterest
Years Experience: 20+
Domains:
- • Cloud Architecture
- • Platform Engineering
- • Network Architecture & Capacity
- • Agentic Infrastructure
Tech Stack:
- • AWS (EC2, EKS, ECS, Kinesis)
- • Terraform
- • Kubernetes
- • Python, Go, Bash
Expertise:
- • Multi-region AWS architecture
- • Cloud networking, IP address management, and capacity planning
- • Platform engineering and developer experience
- • Agent orchestration, verification, and local model serving
- • Cost optimization at scale
- • Infrastructure as Code and automation
Notable Achievements:
- • Built open-source infrastructure for reliable human and AI-agent workflows
- • Multi-million dollar annual infrastructure cost savings
- • Reduced a critical DR recovery path from 20 days to 8 hours
- • Built enterprise DR at a fraction of vendor cost
- • Led multi-region expansion across 4 AWS regions
- • Mentored engineers across infrastructure, security, and reliability teams
Let's Connect
Interested in AI infrastructure, platform engineering, or just want to connect? I'm always happy to talk shop.
Get in Touch →