Skip to content
View alvindcastro's full-sized avatar

Block or report alvindcastro

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
alvindcastro/README.md

Alvin De Castro

Backend-focused AI automation engineer building practical, testable AI systems in Go.

I work on AI products that connect LLMs to real workflows: RAG, agents, workflow automation, evals, observability, and privacy-aware integrations. My focus is not just making a chatbot respond, but making a system that can be tested, audited, deployed, and improved.

Featured AI Projects

Project Domain AI Pattern Backend Evidence
Wiremind Cybersecurity Multi-agent investigation Go, Python, LangGraph, Claude, Redis, Postgres, ChromaDB PCAP metrics (11,562-packet malware capture), MITRE ATT&CK-mapped reports, observability (Prometheus, Grafana, Jaeger, Loki)
GroupScout Sales intelligence AI enrichment + automation Go, Postgres/pgvector, n8n, Claude/Ollama, Slack, Resend Scraped sources, scoring, alerts, promptfoo evals, Prometheus/Grafana/Loki
AskOC MVP Higher ed RAG + workflow automation Go, mock enterprise APIs (Banner, CRM, payment, LMS), Docker Compose Evals, privacy boundary, smoke tests

Work at Douglas College

Internal systems (private repos).

  • Ask Banner (CiteSearch): Go RAG service for Banner ERP and SOP questions, using Azure OpenAI (GPT-4o-mini) and Azure AI Search hybrid vector + keyword retrieval. Multi-source ingestion (SharePoint, file uploads, web, Banner DB, helpdesk tickets) with deterministic chunk IDs and idempotent upserts. Botpress chat front end.
  • QA Reporter: Java/Spring Boot, Playwright, and Allure test automation for Banner, with GPT-4o-mini generating audience-specific summaries for management, module business owners, and developers. Reports delivered through Azure Logic Apps to Blob Storage and SharePoint.

What I Build

  • AI workflow systems that connect LLMs to real business processes
  • RAG systems over approved knowledge sources
  • Agentic pipelines with clear boundaries and human handoff
  • Go services with tests, traces, logs, and deployment-minded structure
  • Practical AI tools for education, security, sales intelligence, and operations

Current Focus

I am exploring local and hybrid AI systems using Go, Qwen/Ollama, llama.cpp, OpenTelemetry, eval gates, and workflow automation. I am especially interested in AI systems that are useful, bounded, observable, and safe to operate.

Tech I Use

Go, Java, Spring Boot, TypeScript, React, Python, PostgreSQL, Redis, Docker, OpenTelemetry, Playwright, Claude, Ollama, LangGraph, pgvector, Prometheus, Grafana, Azure OpenAI, Azure AI Search, Azure Logic Apps, n8n, ChromaDB, llama.cpp, Jaeger, Loki, Kubernetes, GitHub Actions.

Certifications

Microsoft Certified: Azure DevOps Engineer Expert (AZ-400), Azure Administrator Associate (AZ-104), Azure Fundamentals (AZ-900).

Writing

alvindcastro.github.io

Pinned Loading

  1. askoc-mvp askoc-mvp Public

    Go

  2. groupscout groupscout Public

    Go

  3. wiremind wiremind Public

    Go

  4. citesearch citesearch Public

    Go

  5. otel-local-agent otel-local-agent Public

  6. sf-mvp sf-mvp Public

    Go