CHRISTIAN MACION AI Engineer · Multi-Agent LLM Systems · Eval-Harness & Cost-Aware LLM Deployment christianmacion26@gmail.com · +63-991-616-2630 · Digos City, Davao del Sur, Philippines (UTC+8) · LinkedIn === FIT FOR: AI Engineer · LLM Application Engineer · Eval Engineer · Agent Reliability · Applied AI : Anthropic · OpenAI · Scale AI · Surge === SUMMARY I build production multi-agent LLM systems that turn multi-day expert work into reliable, cost-aware automation : and keep them trustworthy with evaluation harnesses, model-routing cost discipline, and honest validation. I designed one reusable architecture from the ground up (no framework) and shipped it across four systems (quant research, venture incubation, content, knowledge base); I authored the curriculum that teaches others to optimize their own. In the most intensive engagement, I operated an 11-agent orchestrator-worker research platform with contracted hand-offs, separation-of-duties between generation and validation agents, a 31-gate statistical evaluation harness implemented scipy-free in numpy, a tiered model-routing layer (Opus / Sonnet / Haiku / Fable) with per-dispatch token budgeting, structured-output contracts gated by a mechanical validator, persistent agent memory with provenance governance, and ~76,500 LOC of light-dependency Python supporting it : plus a telemetry dashboard reading LLM session logs and an offline notebook server. Specialist in the scarce 2026 skills: eval design, agent reliability, and LLM cost modeling. Strong command of multi-agent orchestration (orchestrator-worker / supervisor-worker), tool use / function calling, MCP, Claude Agent SDK / Claude Code, RAG with measured retrieval quality, context engineering, prompt caching, observability, and structured-output validation. Comfortable across Python (numpy, pandas, pyarrow), SQL, AWS S3, Docker, and PDF/PPTX/DOCX generation. I build with large language models; I do not train them. Before the AI work, a self-directed quant-trading practice (crypto and futures) that gave me the adversarial-validation discipline underpinning every eval gate I write. Available 30 hrs/wk for an AI / agent engineer role. CORE SKILLS • Multi-Agent Orchestration (orchestrator-worker / supervisor-worker) • LLM Evaluation Harness Design (statistical-gate validation) • Model Routing & Token-Cost Optimization • Context Engineering • Structured-Output Contracts & Validation (must exit 0) • Tool Use / Function Calling • Hand-off Contracts & Context Isolation • Model Context Protocol (MCP) • Claude Agent SDK / Claude Code • LLM-as-Judge Pipelines • Adversarial Verification & Refutation • Regression / Holdout Discipline • Persistent Agent Memory & Provenance Governance • Prompt Caching • RAG / Chunking / Source-Grounded Generation • Retrieval Scorecards (recall@K, MRR@K) • Observability & Telemetry Dashboards • Python (numpy, pandas, pyarrow, boto3) • SQL • AWS S3 Data Lakes • Docker (working) • JavaScript • PDF / PPTX / DOCX Generation • Headless-Chrome Rendering • AI-Assisted Research & Paper Summarization • Statistical & Mathematical Modeling (foundational) • AI Tools (ChatGPT, Gemini, Claude) • Research Documentation & Communication EDUCATION • Financial Management Studies (current) — University of Mindanao (UM) (1st Semester, AY 2026-27 (enrolled)) • Engineering Studies (units completed) — University of Southeastern Philippines (USeP) (2022 - 2024) • Philippine Science High School (PSHS) — Philippine Science High School System : Southern Mindanao Campus (Graduated 2022) • Certified Technical Analyst Program — Society of Technical Analysts of the Philippines (STA Tier-1) (Dec 2025) TECHNICAL TOOLS • Python (numpy / pandas / pyarrow / boto3) • JavaScript • SQL • AWS S3 • Docker (working) • Google Colab / Jupyter • Git & GitHub • Notion • ChatGPT / Claude / Gemini APIs • Claude Agent SDK / MCP • Headless-Chrome Rendering • Microsoft Excel / Google Sheets RELEVANT EXPERIENCE AI SYSTEMS ENGINEER · LLM RESEARCH PLATFORM (CONTRACT) — SYSTEMATIC-STRATEGY DESK (NDA-PROTECTED; CLOSED PAST CONTRACT 03/2026 - 06/2026) Remote · 03/2026 - 06/2026 • Architected and operated an 11-agent orchestrator-worker AI research platform (~27,500 words of role-scoped agent charters) with contracted hand-off packets, few-shot routing, separation-of-duties between generation and validation agents, and a scribe-owned single source of truth. • Designed a 31-gate statistical evaluation harness (block-bootstrap CIs, random-timing nulls, walk-forward, 5-era stability, Deflated Sharpe, CSCV-based PBO, Minimum Backtest Length) implemented scipy-free in numpy : kills false results before they cost anything. • Built a tiered model-routing policy (Opus = judgment / Sonnet = assembly / Haiku = mechanical / Fable = hardest autonomous) with per-dispatch token budgeting : the cost layer that keeps a multi-agent system (measured at ~15× single-agent token cost) economical to run continuously. • Engineered structured-output contracts gated by a mechanical validator (must exit 0), persistent file-based agent memory (59 typed, cross-linked Markdown records with provenance governance), and ~76,500 LOC of light-dependency Python including a telemetry dashboard reading LLM session logs. • Caught and documented 3 of my own false positives in a single session : banking each as enforced methodology : and authored an institutional operating standard and 16 research studies adopted across the desk. FOUNDER & AI SYSTEMS ENGINEER — MACION VENTURES Remote · 2026 - Present • Built a 7-agent venture-incubation pipeline (5 judgment-tier + 2 mechanical) with 10 lifecycle skills; produced 31 decision-grade artifacts and engineered an anti-self-approval governance pattern (the agent that proposes never approves). • Encoded Philippine tax/regulatory rules (DTI/SEC/BIR/LGU, ₱3M VAT threshold, 8%-flat vs graduated election) directly into agent and skill prompts. AI SYSTEMS ENGINEER (INDEPENDENT) — EDITORIAL / CONTENT AUTOMATION Remote · 2026 - Present • Built an 8-agent content-production pipeline (topic-scout → researcher → drafter → editor → producer → art-director → chart-maker → exporter) and a 290-line AI-output (slop) evaluator scoring drafts on 13 literature-grounded metrics; drove a real draft from HEAVY (index 81) to CLEAN (index 3). • Engineered a dependency-free rendering pipeline (HTML/SVG → headless-Chrome PNG; Markdown → publish-ready PDF). TRADING PLATFORM TESTING & AI-WORKFLOW RESEARCH ANALYST (CONTRACT) — AI-ENGINEERING CONSULTANCY (CLOSED PAST CONTRACT) Remote · 04/2026 - 06/2026 • Conducted structured testing of trading platform onboarding flows to evaluate execution clarity, workflow logic, and system usability from a trader-first perspective. • Applied AI prompting tools (ChatGPT, Claude, Gemini) to accelerate literature review and pattern recognition while manually verifying accuracy of all outputs. • Translated execution observations into structured insights to support product and trading workflow optimization. INDEPENDENT AI SYSTEMS ENGINEER — PERSONAL PORTFOLIO & SELF-DIRECTED STUDY Remote · 01/2025 - 05/2026 • Designed and shipped multi-agent LLM systems end-to-end: agent charters, eval harnesses, model-routing policy, structured-output contracts, persistent memory, and Python tooling : for content automation, venture incubation, and personal-knowledge workflows. • Built, audited, and open-sourced a portfolio of runnable AI projects (RAG scorecard, ReAct tool-calling agent with OTel traces, LLM-as-judge validated vs humans, MCP eval server, self-critiquing reflection agent, AI-slop evaluation gate). FINANCIAL MARKET EDUCATOR & AI INTEGRATION SPECIALIST (ONGOING SIDE) — INDEPENDENT / VARIOUS UNIVERSITIES AND COMMUNITIES Remote / Davao Region · 12/2024 - Ongoing • Delivered workshops, webinars, and presentations on financial markets, AI applications, and emerging technologies. • Served as a guest speaker at universities and professional communities on topics involving AI, financial literacy, and blockchain technologies. ACADEMIC & PERSONAL PROJECTS • Multi-Agent AI Architecture (Field Guide) (available on request) — Curriculum authored from first principles (no framework) covering orchestrator-worker topologies, eval-harness design, model-routing policy, structured-output contracts, persistent agent memory, and cost discipline. Adopted as the teaching reference for a research-desk engagement. • RAG Retrieval Scorecard (rag-recall) (available on request) — Measured RAG with a retrieval scorecard: recall@3 = 0.886, MRR@3 = 0.805. Supports recall@K / MRR@K evaluation across chunking strategies and vector-store choices. • LLM-as-Judge Validated vs Humans (available on request) — Judge harness validated vs human raters: Cohen's κ = 0.58, 95% bootstrap CI, position-bias measured. CERTIFICATIONS AI, DATA SCIENCE & TECHNOLOGY • Introduction to AI Agents : DataCamp • AI Fundamentals : DataCamp • Artificial Intelligence Fundamentals : IBM • Introduction to AI : Google • AI for the Modern Workforce : Ateneo de Davao University FINANCE, TRADING & ECONOMICS • Certified Technical Analyst Program : Society of Technical Analysts of the Philippines (Tier-1) • Financial Trading in Python : DataCamp • Math for Finance Professionals : DataCamp • Excel for Financial Planning and Analysis (FP&A) : LinkedIn • Foundations of Growth Equity : Goldman Sachs • JPMorgan Chase Investment Banking Job Simulation : Forage MATHEMATICS, STATISTICS & PROGRAMMING • Getting Started with Python for Finance : LinkedIn • Python Essentials 1 : Cisco Networking Academy • SQL for Finance Professionals : LinkedIn • Understanding Data Science : DataCamp SELECTED EVENTS, HACKATHONS & PROGRAMS • Galactic Problem Solver : NASA Space Apps Challenge (Zurich, CH) : Oct 2025 • BIDA META AICCELERATE 2025 (5-day intensive) : Meta & Bayan Academy : Dec 2025 • UP Data Science Society : Alternative Classroom Learning Experience 2026 : Feb 2026 • LimitlessBiz: AIM ASEAN Training Programme : Limitless Lab : Mar 2026 • Guest Speaker : USeP CBA Annual Business Expo 2026 : Feb 2026 • Sui Builders Program Davao : Sui Foundation : Apr 2026