tag:christianmacion26.github.io,2026:feed:all Newest projects, research artifacts, and client solutions. Quant, AI, OSS. 2026-07-09T15:41:16.774Z Christian T. Macion christianmacion26@gmail.com https://christianmacion26.github.io/ tag:christianmacion26.github.io,2026:solution:11-agent-eval-platform 2026-07-09T15:41:16.774Z

A small systematic-trading desk needed an AI workflow that could keep pace with the research pipeline *without* shipping hallucinated or unverifiable analysis. The naive path — a single agent with one prompt — produced plausible but unfalsifiable outputs. The desk needed something closer to a research organization than a chatbot.

tag:christianmacion26.github.io,2026:solution:7-agent-venture-pipeline 2026-07-09T15:41:16.774Z

An operator-led venture incubation arm needed a repeatable way to triage, brief, and ledger new business ideas — without one human bottleneck, and without the agent that proposes a decision being the same one that approves it.

tag:christianmacion26.github.io,2026:solution:8-agent-editorial-pipeline 2026-07-09T15:41:16.774Z

A high-volume editorial workflow was generating content with a measurable "slop index" — generic, templated, easily-detected text. Quality gate was after-the-fact and manual. Production scaled faster than the editorial team could review.

tag:christianmacion26.github.io,2026:solution:9-project-quant-library 2026-07-09T15:41:16.774Z

Most online quant research demos are not reproducible: closed datasets, undisclosed parameters, un-reported multiple-testing bias, and no OOS discipline. A hiring-grade research portfolio needs every one of those addressed explicitly.

tag:christianmacion26.github.io,2026:solution:crypto-stat-arb-funding 2026-07-09T15:41:16.774Z

Crypto perps run funding payments every 8h. A naive long-short book ignores funding carry and bleeds slowly when the spread is inverted. A working stat-arb needs the funding cost added back to P&L *before* sizing.

tag:christianmacion26.github.io,2026:solution:cta-curriculum-public 2026-07-09T15:41:16.774Z

The Philippines has limited access to rigorous CTA-grade technical analysis training. Most curriculum is either imported (US/UK, expensive) or shallow (TA-by-rote). There's a gap for affordable, rigorous, evidence-based technical analysis education.

tag:christianmacion26.github.io,2026:solution:deflated-sharpe-gate 2026-07-09T15:41:16.774Z

Naive Sharpe ratios ignore the search effort — if you try 100 variants of an idea, the best-looking one will overstate the true edge. The desk needed this multiple-testing correction built into the validation pipeline, not stapled on at the end.

tag:christianmacion26.github.io,2026:solution:eval-mcp-server 2026-07-09T15:41:16.774Z

LLM eval harnesses live in code or in spreadsheets — neither is a clean integration target for multi-agent pipelines. A multi-agent system needs the eval gates exposed as **tools**, not as Python imports.

tag:christianmacion26.github.io,2026:solution:lookahead-bias-audit 2026-07-09T15:41:16.774Z

Look-ahead bias is silent: a backtest that uses future data looks great until you deploy it. A serious research portfolio needs an **explicit audit suite** — a checklist of failure modes with mechanical tests.

tag:christianmacion26.github.io,2026:solution:txn-cost-aware-backtest 2026-07-09T15:41:16.774Z

Backtests that ignore transaction costs, slippage, and latency are the most common source of overfit research. A backtest engine needs all three treated as first-class inputs, not as after-the-fact deductions.

tag:christianmacion26.github.io,2026:project:quant/09-lookahead-bias-audit 2026-06-18T00:00:00.000Z

A leak inflates Sharpe 0.59 → 5.07; the shift test exposes it as 88% phantom.

tag:christianmacion26.github.io,2026:project:quant/08-backtest-engine-costs 2026-05-28T00:00:00.000Z

High-turnover signal wins gross (0.20 > 0.13) but loses net of cost — break-even 20 bps.

tag:christianmacion26.github.io,2026:project:quant/07-macro-regime-overlay 2026-05-10T00:00:00.000Z

Vol-managed exposure: Sharpe 0.64 → 0.67, max DD −58% → −42%.

tag:christianmacion26.github.io,2026:project:quant/06-funding-carry 2026-04-05T00:00:00.000Z

Funding +11.9% annualized premium; fade IS 1.11 → OOS −0.05 (decayed post-2023).

tag:christianmacion26.github.io,2026:project:ai/06-slop-scanner 2026-03-22T00:00:00.000Z

13-metric literature-grounded AI-output quality gate — drove a real draft from HEAVY (81) to CLEAN (3).

tag:christianmacion26.github.io,2026:project:ai/05-reflect-revise 2026-03-01T00:00:00.000Z

Reflection-loop agent — mean SLOP 127.5 → 14.0 across drafts, 3/4 improved, 1/4 honest no-progress halt.

tag:christianmacion26.github.io,2026:project:quant/05-pairs-cointegration 2026-02-18T00:00:00.000Z

ADF −2.11 (not cointegrated), half-life 208 d — the fade loses, as the test predicts.

tag:christianmacion26.github.io,2026:project:ai/04-eval-mcp-server 2026-02-04T00:00:00.000Z

MCP server exposing the slop-evaluation gate over Tools, Resources, and Prompts — 20/20 conformance, 100% round-trip parity.

tag:christianmacion26.github.io,2026:project:quant/04-variance-risk-premium 2026-01-25T00:00:00.000Z

Implied > realized 85% of 36 yrs; predicts returns, Newey-West t = +6.5.

tag:christianmacion26.github.io,2026:project:ai/03-judge-harness 2026-01-12T00:00:00.000Z

LLM-as-judge pipeline validated against human raters — Cohen's κ = 0.58 with bootstrap CI and position-bias measured.

tag:christianmacion26.github.io,2026:project:quant/03-timeseries-momentum-voltarget 2025-12-20T00:00:00.000Z

Vol-targeting lifts Sharpe 0.27 → 0.39 and halves max DD (−62% → −30%).

tag:christianmacion26.github.io,2026:project:ai/02-toolcall-agent 2025-12-08T00:00:00.000Z

ReAct-style tool-calling agent with OTel traces, fault injection, and 100% tool/arg correctness.

tag:christianmacion26.github.io,2026:project:ai/01-rag-recall 2025-11-15T00:00:00.000Z

RAG service that proves its own retrieval — recall@3 = 0.886, MRR@3 = 0.805 with offline stdlib TF-IDF retriever.

tag:christianmacion26.github.io,2026:project:quant/02-cross-sectional-momentum 2025-10-15T00:00:00.000Z

IS Sharpe 0.91 → OOS −0.03 — an honest decay; bootstrap CI straddles zero.

tag:christianmacion26.github.io,2026:project:quant/01-deflated-sharpe 2025-09-20T00:00:00.000Z

Best-of-160 BTC rule: IS Sharpe 1.14 is only 1.24× the pure-noise expectation — DSR = 0.70 (fail).