Engineering
How we build Dosu: pipelines, evals, infrastructure, and lessons from running AI agents in production.10 posts
How to design a CLI login flow for coding agents
Design a CLI login for coding agents with a ticket flow that pauses, resumes in a new process, and returns the next command in JSON. Learn how Dosu built it.
Taylor DolezalHead of OSSAgentic Infrastructure: What We Learned Hosting a Panel With LanceDB, CodeRabbit, and CrewAI
Four companies focused on agentifying their tech stack. One rewrote its codebase for AI. Another found PII leaking through OpenTelemetry traces.
Word2Vec To Transformers And Where We Are Today With AI
Understanding the foundational technology behind Dosu: from Word2Vec embeddings to Transformers and modern LLMs.
George XuFounding BizOps LeadFrom Celery to DBOS: Scaling Dosu’s Pipelines to 20k Workflows per Hour
See how Dosu scaled AI pipelines with DBOS: from Celery to durable workflows, real-time observability, and 20k jobs per hour serving 50k+ projects
Marcos PlaconaDevRelHow LanceDB Supercharged Our Knowledge Graph
How we scaled our knowledge graph by switching from pgvector to LanceDB, enabling millisecond search across millions of vectors.
Marcos PlaconaDevRelBetter Agent Testing with Distributed Dependency Injection
How Dosu uses distributed dependency injection to improve agent testing and reliability.
Michael MangusHead of ProductCelery Preserializers: A low-friction path to Pydantic support
Learn how to add Pydantic support to Celery tasks with minimal friction using preserializers.
Michael MangusHead of ProductHow Dosu Used LangSmith to Achieve a 30% Accuracy Improvement with No Prompt Engineering
There are many techniques for “teaching” LLMs to improve their performance on specific tasks. The most common are:
Devin SteinFounder and CEOIterating Towards LLM Reliability with Evaluation Driven Development
How to improve LLM reliability through evaluation-driven development practices.
Devin SteinFounder and CEOReliable Data Orchestration for AI Applications
How to build reliable data orchestration systems for AI applications.
Devin SteinFounder and CEO