Skip to content
View coder058's full-sized avatar

Block or report coder058

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
coder058/README.md

Jordi Lluis

Amsterdam. I like tackling everyday problems that used to take too long, with the tools that are actually useful now. Then I check whether what I built works.

Portfolio · Résumé · Engineering notes

Live demos: Pattern Forge · Relay · Python trading bot archive

Pattern Forge

CI

Result: replay without look-ahead; CI runs tests, Docker and a PostgreSQL restart. The public demo does not serve that database.

Architecture: browser → Next.js candle API → Hyperliquid public quotes. A separate WebSocket carries live mid-price and does not write into replay. Recorded replay uses only the selected prefix; incomplete higher-timeframe groups are omitted, not guessed. Persistence is for local use and CI.

What I built: React/TypeScript workspace, validated candle snapshots, Docker build checks and failure-handling tests.

Code: app/ · closed-candle tests · ingest

npm ci && npm test && npm run build

Relay

CI

Architecture: React UI → stateless FastAPI → one deterministic matcher, also exposed as four read-only MCP tools. Returns quoted sentences, not a hiring score. Paste is the real input; Arbeitnow is a bounded demo snapshot.

What I built: search_job_board, summarize_job_board, review_job_evidence, export_job_review. Duplicates keep changed wording.

Code: backend/ · evidence tests · MCP

Matching is literal. An agent that calls these tools can still summarise; Relay does not decide eligibility.

Python trading bot

Decision: publish the loss and keep branch prices distinct. Early tickets were often $0.96–$0.99; StratA used $0.40–$0.72 with 11–15 seconds remaining; UC looked at leftover $0.01–$0.20 asks.

What is public: ledger script, timing parser, browser case study. The private bot is not in this repository. Live trading stopped.

Code: analyze.py · EVIDENCE.md

An API acknowledgement is not a fill. Dublin was not compared with another region.

Open source

Merged Haystack #12635: AutoMergingRetriever now treats 0 as a stored hierarchy value, not missing metadata. That is one accepted fix, not Haystack product work.

City Gardens (team, 2023)

Le Wagon Barcelona. Rails and PostgreSQL. Not sole-authored and not maintained.

My merged pull requests: parcel reservations, add-event button, logout, styles, fontsize. I also contributed event-index partials, a search bar and Cloudinary for event photos. That is not the whole schema.

Schema

Other public exercises

DispatchOps · Transcript Desk · Wikinimous · API labs · Rails labs

Pinned Loading

  1. relay relay Public

    Read-only job evidence desk over HTTP and MCP: exact source quotes, duplicate retention and portable reviews. Deterministic, not AI eligibility scoring.

    Python

  2. pattern-forge pattern-forge Public

    React market workspace: live public quotes, a validated candle API and recorded replay. Docker build and failure-handling tests.

    TypeScript

  3. polybow-case-study polybow-case-study Public

    Reproducible live-trading ledger postmortem, anonymized data and browser archive. Includes the full loss; no profitability claim.

    JavaScript

  4. coder058 coder058 Public

    Portfolio README: selected projects, data tables, resume packs, apply packet.

  5. dispatchops dispatchops Public

    Interactive React and TypeScript dispatch simulator with an explainable, human-approved recommendation flow.

    TypeScript

  6. profile profile Public

    Selected project stories and live demos by Jordi Lluis.

    HTML