AI Systems Portfolio · 2026

Daria
Gavrilova

AI Content Strategist & Systems Architect I build AI systems that organisations actually adopt — not just systems that technically work. My background spans 15+ years of investigative journalism, 7 years of curriculum design, and a full-stack AI implementation practice. The combination is not accidental.
75–80%Workflow
Automated
85%KPI Success
Rate
4Working
Languages
Scroll
01 —

Most AI systems fail not because
they don't work — but because
nobody uses them.

02 —

Selected Work

Enterprise AI SystemFeb — Apr 2026

Enterprise Content
Intelligence Platform

80%
of manual proposal
work automated

Enterprise sales reps were spending the majority of their time manually assembling client proposals and course outlines from fragmented course content — a process that was slow, inconsistent, and impossible to scale. No tooling existed to support this.

I designed and built a full-stack content intelligence platform: a structured GitHub repository as the canonical content source, a React/Vite frontend that reads from GitHub and collects enterprise client parameters, and a Claude API integration layer governed by the atomic modules logic. The system reduced manual workload by 80%, allowing to increase the client flow fivefold.

01 / Source
Repository
GitHub
Content · Meta · Decks
02 / Interface
Frontend
React / Vite App
Enterprise Config Layer
03 / Intelligence
AI Layer
Claude API
+ atomic modules
04 / Output
Artifacts
Proposals · Outlines
PDF · Markdown · HTML

"I have waited half a year to see something like this."

— CEO, live product review session

Stack
React / ViteClaude APIGitHubCursorMarkdownHTML Deck SystemGENERATION_RULES.mdAtomic Content Architecture
Editorial AI System · Consultancy2025 — Ongoing

Editorial Knowledge
Automation System

17+
episodes
processed

The client — an investigative podcast brand with a published archive of 17+ episodes — was producing high-value interview content that disappeared into static audio files. No searchable knowledge base. No content reuse. No compounding value. Each episode existed in isolation. Significant editorial intelligence was being generated and then lost.

I designed and built a three-tool automation stack (Whisper AI → Cursor → GitHub) and a four-type knowledge atom model to make every interview compound. Each transcript is processed into structured atoms: Voices (direct source quotes), Patterns (recurring themes across guests), Tensions (contradictions worth investigating), and Leads (story threads for future episodes). The GitHub repository functions as a living editorial brain — queryable, versioned, and directly connected to downstream distribution.

Click any node to explore · Watch the data flow
Voices
Guest quotes · by theme
Direct quotes from guests, tagged by theme. The atomic unit of the archive — one voice can appear across multiple outputs without duplication.
Patterns
Recurring · across guests
Things multiple changemakers keep saying independently. The most editorially powerful atom — signals a structural truth worth a dedicated essay or episode.
Tensions
Contradictions · to explore
Contradictions between guests on the same topic. Tensions are the engine of future editorial — the library as a machine for finding the next story.
Leads
Future episode · threads
Threads that deserve a full conversation but haven't had one yet. The library as a living commissioning document — every interview generates the next one.
4 types
Knowledge atom taxonomy — Voices, Patterns, Tensions, Leads — designed for editorial intelligence, not generic tagging
17 eps
Full archive processed into structured Markdown with name-correction pipeline and replacement logging
∞ reuse
Every interview permanently queryable — a compounding editorial asset, not a static audio file
Stack
Whisper AICursorClaude APIGitHubPythonJSON Schemaatoms.jsonKnowledge Architecture
03 —

Background

Certification · 2025
Líders Digitals AI Bootcamp
Completed with distinction — allWomen × Generalitat de Catalunya. Covers AI strategy, automation, no-code systems, and LLM implementation.
Certification · 2016
Linguistics — Universiteit Leiden
Foundations of Linguistics via Coursera. Language structure, semantics, pragmatics. Directly informs how I approach prompt architecture and LLM behaviour.
Teaching · 2019–2026
AI Curriculum Design & Delivery
500+ students trained at UAB in applied AI and marketing strategy. Courses designed from scratch, including "Digital Copywriting and AI" — among the first of its kind at the institution.
Tools & Stack
Technical Competencies
React / Vite · Claude API · GitHub · Cursor · Whisper AI · Zapier · N8N · Lovable · Markdown systems · JSON Schema · Prompt architecture · No-code automation