Skip to content
View PhelipeHonesto's full-sized avatar

Block or report PhelipeHonesto

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. Markdown is supported. 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
PhelipeHonesto/README.md

P.H. · Aviation intelligence systems

Phelipe Honesto (PhelipeHonesto) builds AI agent systems and data infrastructure for private aviation: flight-data ingestion, entity resolution over operators and aircraft, charter-brokerage orchestration, and memory for the agents that run it. The core stack is Dumont Agent, OSA Data Platform, and ph-brain.

Tip

Former airline first officer turned software builder. 2,000+ flight hours, 1,000+ airline flights, now shipping spec-first, test-heavy systems for the industry I flew in. Built under Orange Sunshine Aviation · Writing at Loud Hawk

Core stack · Flagship systems · How I build · Start here · Background


How the stack fits together

OSA Data Platform ingests and resolves. ph-brain remembers. Dumont reasons and acts. Loud Hawk publishes what the data says.

flowchart LR
    A[ADS-B feeds] --> D
    B[FAA registry] --> D
    C[NTSB records] --> D
    D[OSA Data Platform<br/>ingestion · entity resolution · analytics] --> E[(Postgres / pgvector)]
    E --> F[Dumont Agent<br/>charter orchestration]
    E --> G[Loud Hawk<br/>aviation intelligence]
    H[ph-brain<br/>memory substrate] <--> F
    F --> I[Quotes · operator sourcing · close]
Loading
Layer System Responsibility
Ingestion OSA Data Platform Pull ADS-B, FAA registry, and NTSB data into one schema-validated store
Entity resolution OSA Data Platform Collapse messy records into canonical operators, aircraft, and tail numbers
Memory ph-brain Structured ingestion of documents, notes, and media into an embedded knowledge substrate, served over MCP
Agent control Dumont Agent Run the charter pipeline with validation gates at every stage and learn from closed deals
Publishing Aviation News Aggregator → Loud Hawk Turn industry news and data signals into an aviation intelligence newsletter

Flagship systems

Project What it is Evidence Stack
dumont-agent AI charter-brokerage orchestrator: intake → validation → quoting → operator sourcing → close, with a learning loop over closed deals 681 tests · 22/25 end-to-end scenarios · auth enforced on every private route · MCP server (17 tools) · CI TypeScript, Vitest, Zod, Postgres, MCP
osa-data-platform Aviation data platform: ADS-B / FAA / NTSB ingestion, canonical entity resolution, analytics compute 21M flight legs · 484 tests · LLM resolution eval: 0 false matches, 0 fabricated IDs · CI TypeScript, Supabase / Postgres, pgvector
ph-brain-engine Personal intelligence substrate: embeds a markdown vault and serves hybrid semantic + keyword search to agents over MCP 22 tests · CI · retrieval core ported from GBrain (MIT) TypeScript, Bun, Voyage AI, pgvector, MCP
Aviation-News-Aggregator Loud Hawk v2: 74-source aviation catalog (25 active by default), 14 fetch strategies, per-source circuit breakers, authority-weighted ranking API smoke tests · CI · Docker Python

How I build

Flight decks run on checklists, so my code does too.

Practice What it means in the repo
Spec first Every system ships with a written specification before code (SPARC-style)
Test heavy 1,100+ tests across the stack; validation runs at every pipeline stage, not only at the end
Schema at the edges Zod validation on every ingest and every agent handoff
Evaluation over vibes LLM steps get eval harnesses and scored passes before they're trusted
Ship small, ship visibly CI on every push, containerized deploys, documented dev workflows

Start here

Goal Project
See a multi-stage AI agent with validation gates dumont-agent
Ingest and resolve FAA / ADS-B / NTSB data osa-data-platform
Build an MCP-served personal knowledge base ph-brain-engine
Follow the aviation intelligence newsletter Loud Hawk

Toolbox

TypeScript Python PostgreSQL Supabase React Vite Vitest Docker GitHub Actions Railway MCP Obsidian

Background

  • Airline First Officer, SkyWest Airlines (E175): 1,000+ flights, 76,000+ passengers, 100% procedural compliance in an FAA Part 121 environment
  • Flight Instructor (CFI / CFII / MEI): trained 30+ pilots to a 90%+ checkride pass rate
  • Jump Pilot: PIC for a 65-way formation skydiving world record (Nov 2022)
  • Before aviation: ~3 years running restaurant and hospitality operations in Denver
  • Brazilian-American with family roots at Embraer in São José dos Campos · 🇺🇸 🇧🇷 · English / Português / Español

Based in Denver. Open to conversations about aviation data, agent systems, and charter tech: phelipe.hon@gmail.com


Built in public so the work can be inspected, tested, and dated. ✈️

Pinned Loading

  1. ph-brain-engine ph-brain-engine Public

    Personal intelligence substrate: embeds a markdown vault into pgvector and serves hybrid search to agents over MCP.

    TypeScript

  2. osa-data-platform osa-data-platform Public

    Aviation data platform: ADS-B, FAA registry, and NTSB ingestion with canonical entity resolution for operators and aircraft. 490+ tests.

    TypeScript

  3. dumont-agent dumont-agent Public

    AI charter-brokerage orchestrator: intake → validation → quoting → operator sourcing → close. Guardian gates every stage. 670+ tests, MCP server.

    TypeScript

  4. Aviation-News-Aggregator Aviation-News-Aggregator Public

    Loud Hawk v2: aviation news aggregator. 74-source catalog (25 active by default), per-source circuit breakers, dedup, authority-weighted ranking.

    Python 1