Practical tools built with AI — assessment chatbots, agent skills, and lightweight ML systems — spec-driven and human-gated where it matters.

• citation-audit — audits a finished reference list against the bibliographic record, catching fabricated, misattributed, and misdirected citations — including identifiers that resolve cleanly to an entirely different paper. Keyless; the calls that change what a document claims stay with the human
• agentic-research — 23-skill PRISMA 2020 / GRADE systematic-review pipeline: a question → a defensible synthesis
• research-writer — subagent-orchestrated, citation-checked literature-review drafting

• dost-progress-report — one Markdown file → a Word progress report in the DOST prescribed format (Forms 6 & 7), with nine checks that catch dropped chapters, stale cross-references, and citation gaps
• keynote-builder — a governed, human-gated pipeline that turns a brief into a stage-ready keynote (slide deck + speaker script)
• office-reports — agent skills for institutional document drafting: travel reports, minutes, and a weekly → terminal M&E reporting pipeline, evidence-only with gap flags
• presentation-builder — composable skills for finished presentations: Marp, native PPTX, HTML slides, talk scripts
• research-deck-builder — a research/training module → a presenter-ready, source-checked deck
• agentic-vault — an Obsidian vault template for working with AI agents (PARA + multi-agent governance)

• aiaas-marketability — a Demand Viability Index chatbot for a localized AIaaS platform (DOST-NAIRA)
• acabai-ph — the DOST-ASTI showcase site for the Philippines' national AI initiatives (NAIRA, DIMER, iTANONG), with an embedded chatbot and automatic failover
• ai-readiness-assessment — an AI-readiness assessment chatbot for Philippine government agencies & NGOs

• inference-bench — an MLPerf + TensorRT inference and GPU/CPU hardware benchmark suite across heterogeneous hardware (RTX 5070 Ti, T4, A100/H200, CPU), with reproducible scripts, notebooks, and results
• benchmarking-harness — a model-class-aware, three-tier evaluation gate for computer-vision models (capability → domain stress → operational safety) with a human review gate and auto-generated Model Cards
• mlops-lite — a full-lifecycle, single-GPU MLOps platform (data → train → serve → monitor → retrain)
• notebooks — ready-to-run Colab speech & ML experiments (ASR, diarization, TTS)

• pdf-signer-pwa — private, offline PKCS#12 PDF signing that never leaves your device