Dmitri Knapp
Professional Summary
Applied AI product leader who builds the systems, not just the strategy — 7+ years turning enterprise design leadership into shipped AI product across B2B SaaS, financial, and federal platforms. At a 250-person retail-tech firm, architected the company's first AI knowledge platform from zero with CEO sponsorship — a 9-layer, access-controlled system now spanning 18 vaults and 23,500+ documents and powering a fleet of 13 agents and 10 reusable skills — then shipped a design pipeline on top of it that produced 456 documented UI states, every answer and screen traceable to a source. Currently co-founding Atelos, a local-first AI knowledge product taken from concept to MVP in ~2 months (owning product strategy while shipping the Electron/React build), and previously solo-built Focus Agents, a multi-agent research-synthesis SaaS that cut 26-hour jobs to 33 minutes. Earlier led design for Kroger's Alternative Profit business ($1.35B FY2024 operating profit), originating the future-state designs for initiatives that secured $35M+ in funding. Specializes in agentic AI systems, LLM-powered products, and grounding product decisions in real data rather than opinion.
Core Competencies
Professional Experience
- Architected the company's first internal AI knowledge platform from zero with CEO sponsorship — a 9-layer, access-controlled system on GitHub unifying company data, meeting transcripts, research, product knowledge, and competitive intel into a single agent-accessible substrate, with PII sanitization, dual-zone executive-authority handling, agent memory (Microsoft Foundry), and tiered branch protection; approved by CEO and IT for company-wide scale, and the CRO called the architecture "brilliant" in writing.
- Built a 13-agent, 10-skill automation fleet (meeting capture, decision logging, world-model sync, doc updates) with 500+ tracked production runs, plus an 8-phase AI-augmented design pipeline on the platform — producing 456 documented UI states and counting toward a ~1,000-screen storefront redesign, all with full source provenance traceable to a sourced problem brief; modernizing the suite with one designer plus agents.
- Compressed design-to-dev handoff from a typical 1–2 week cycle to a single day on a flagship dashboard — turning a stakeholder meeting into developer-ready, edge-case-complete HTML in one afternoon.
- Built repo-based retailer-intelligence scrapers feeding agents real competitive and customer signal — a fleet analysis across the deployed customer base redirected the storefront roadmap away from a feature only ~9% of customers used, eliminating wasted engineering spend before code was written.
- Rolled out the company's first cross-product design system (~30% dev-time reduction) and documented 82 design decisions with source attribution; lead a 2-person design team and partner with Engineering, PM, Sales, and the C-suite on AI strategy and product direction.
- Led design for Kroger's Alternative Profit business — $1.35B in FY2024 operating profit — across credit, gift card, lottery, money services, financial compliance, sports betting and entertainment, and in-store digital media.
- Directed design strategy influencing $35M+ in funded initiatives; repositioned the design team from execution function to strategic decision driver.
- Managed a team of 5 designers and partnered with 15+ product, engineering, and VP-level stakeholder teams across the Alt Profit organization.
- Built 9 enterprise service blueprints mapping 150+ touchpoints to align executives on cross-channel strategy.
- Created governance models and alignment rituals that accelerated cross-team decision-making; maintained 100% team retention through major org restructuring.
- Operationalized AI agent workflows (ChatGPT, n8n, Gamma) reducing research synthesis and executive deck creation from days to minutes.
- Led service design and product strategy for Fortune 500 transformation programs; managed a team of 3 designers.
- Reimagined Realogy's end-to-end agent experience through service blueprints and future-state journey mapping.
- Standardized 30+ consulting frameworks and playbooks adopted across a 70-person UX practice.
- Coached client-facing teams on research, executive storytelling, and stakeholder alignment.
- Directed discovery for a federal data-consolidation platform unifying 1,700+ siloed data sources across a national-scale government environment.
- Service design for the full overhaul of the federal TANF platform — the system administering $16.8B in block grants to state welfare programs.
- Mentored 7 team members in UX research methods and delivery practices.
- Designed responsive dashboards visualizing real-time data from 30,000+ IoT sensors across cold-storage facilities.
- Led pilot releases through research, requirements gathering, and rapid prototyping.
- Standardized usability testing, persona, and design handoff processes.
- Built the complete design system and component library, reducing development time by 50%.
- Defined product vision, monetization strategy, and end-to-end UX flows.
- Led MVP prototyping, usability testing, brand identity, and launch marketing.
Education
Independent AI Work
- Help lead an early-stage startup building a local-first Electron desktop app that wraps the Claude API, pre-assembles context from a deterministic vault structure, and executes packaged "skills" — drove it from concept to MVP in ~2 months.
- Own product end-to-end: ran 10+ primary research studies (jobs-to-be-done synthesis, competitor teardowns of Obsidian, Notion AI, and Logseq/RAG tools), defined a 3-persona ICP, pricing model, and brand system, and co-authored the V1 business case (TAM/SAM, competitive analysis, GTM, financial model).
- Own the full TypeScript/React front-end — app shell, onboarding, chat composer, skill library, projects, and live vault-context visibility — shipping 35+ reviewed PRs through a trunk-based flow.
- Contribute core architecture: authored 8 Architecture Decision Records and a "cascade" write-integrity primitive that routes every vault mutation through a single guarded API, enforced by a CI lint gate.
- Shipped tiered monetization (free/paid entitlements + feature flags) and bring-your-own-key security with Anthropic API keys encrypted at rest via Electron safeStorage.
- Designed and solo-built a production multi-agent SaaS delivering AI research-synthesis agents into users' Google Drive — 3 production agents reduced 26-hour synthesis jobs to 33 minutes (47x speedup) at 4.2–4.5/5.0 composite quality across 12 dimensions.
- Built shared infra, stateless horizontal scaling, Stripe billing, and a 12-dimension quality scorecard. Stack: n8n orchestration, Claude (Sonnet 4.5) API, Gemini for audio, Framer/React frontend, Google Workspace integrations, Zero Data Retention by design.
- Authored an Empathy & Ethics Framework as a core architecture constraint; chose to stop scaling when production behavior validated a misuse concern — a deliberate call on responsible AI deployment.