Justin Castilla

@elastic_devs I built an Elastic MCP server to help plan my commute to and from home, work, and events in my calendar. Reduces my traffic time and eases the daily commute: github.com/justincastilla…

Deep Agents helps coordinate the subagents to perform individual dedicated tasks, Elastic Agent Builder assists with the natural language queries on previous research, and Tavily reaches out to pull down the most up-do-date information. Should be a great time!

We'll build out an app to gather framework sentiment anlysis, repository liveliness metrics, and overall evaluation scores to offer up a "developer advocate" service to those wanting to vet the best tech out there.

Hey everyone! I'll be presenting a workshop on @LangChain 's Deep Agents agentic harness with @elastic agent builder and @tavilyai this friday at the Women in Data Science Conference. Register for the event here: widspugetsound.org

Start today → docs.langchain.com/oss/python/dee…

@CastillaCodes has a Spotlight Session at Nebraska.Code() this July. Find 'Build your own Developer Advocate with DeepAgent!' here: nebraskacode.amegala.com #DeepAgent #Frameworks #Repositories #LangChain #SubAgents #AI #GitHub #Opensource @elastic @Fiserv @omahanehr @StreckLLC @osaic_inc #AI @Centene @KawasakiUSA @ConagraBrands #Tech

@CastillaCodes has a Spotlight Session at Nebraska.Code() this July. Find 'Tracking Longterm Health with a Sympathetic Voice - empowering Agentic AI to actually listen' here: nebraskacode.amegala.com #AgenticAI #MCP #LLMs #Midwest #Python #API @LiveWellNE @UnitedHealthGrp @rightathomeUS @neiaredcross @westcorphealth @NRCHealth @OmahaPython @Bryan_Health @MercyOne_Iowa

A competent RAG pipeline with a billion parameters? It’s no longer just a pipe dream. The recent release of Gemma and Qwen models give developers access to models as small as under 1B parameters that can run on not just your computer, but edge devices. Combine them with Jina AI’s v5 text embedding models, and you can have a competent RAG pipeline for around 1B parameters, or an even better one for only 5-10B parameters. What’s even better is that these Gemma and Qwen models are multimodal, so they can understand images as well as text. You can combine these with Elasticsearch (running on your machine or in the cloud) and you’ll have an efficient RAG pipeline at a fraction of the compute you used to need.

Later this week I'll be at the Women in Data Science Puget Sound conference with a quick workshop in Python for Deep Agents. It's open to all and has a fantastic lineup of talks and workshops. I love conference season! widspugetsound.org

Then it was northward to the Vancouver Cloud Summit to hang out at our booth and act as judge for the Elastic-sponsored Hackathon. It's truly amazing what an enthusiastic and empowered group can create in such little time!

Had an amazing last few days working with the community! On Thursday I had the opportunity to run a workshop with SeattleJS covering LangChain's Deep Agents agentic harness with Tavily and Elastic's Agent Builder to build out a powerful technology research application.

What did you build with Elasticsearch in 2025? Share your creation below!

@elastic_devs I built an Elastic MCP server to help plan my commute to and from home, work, and events in my calendar. Reduces my traffic time and eases the daily commute: github.com/justincastilla…

I've moved over to Bluesky! @justincastilla.bsky.social