I help early-stage startups go from idea to real traction. My role sits somewhere between founder, full-stack engineer, and product owner: I design, build, and iterate on MVPs, while keeping an eye on strategy, metrics, and the real behavior of users.
How I work with startup teams
1. Discovery & alignment – we clarify the problem, the target user, and what "success" means in the next 1–3 months (not "in an ideal world in 5 years").
2. Strategy & SWOT – I run a lightweight SWOT analysis for the product, the tech stack, and sometimes even for the team's skills. This helps to decide what to build first, what to postpone, and where we need extra expertise.
3. Fast MVP cycles – I build a minimal, but real product (not just a prototype) and we iterate in short cycles based on feedback and data.
4. Traction focus – we don't chase "all features at once". Instead, we aim for a clear signal: engaged users, first payments, improved retention, or a validated use case.
Traction over hype
I'm biased towards traction, not decks. Together we define 1–3 simple metrics that matter for your current stage: active users or engaged sessions, first paying customers, or clear time-savings for internal tools. Every sprint, we ask: did we move these numbers – or just ship more code?
Using SWOT for technical choices (example: OCR for invoices)
I also use SWOT not only for business, but for technical decisions. For example, when choosing an OCR engine for invoice processing, I compare options like DeepSeek OCR, Marker, and other engines in terms of strengths, weaknesses, opportunities, and threats.
┌───────────────┬────────────────────────────────────────┬──────────────────────────────────────────┐
│ Engine │ Strengths │ Weaknesses / Risks │
├───────────────┼────────────────────────────────────────┼──────────────────────────────────────────┤
│ DeepSeek OCR │ Strong on complex layouts and │ Newer stack, ecosystem and tooling │
│ │ multi-language documents │ still maturing │
├───────────────┼────────────────────────────────────────┼──────────────────────────────────────────┤
│ Marker │ Good for document structure, │ May require extra post-processing │
│ │ tables, and PDFs │ for noisy scans │
├───────────────┼────────────────────────────────────────┼──────────────────────────────────────────┤
│ Google Docs │ Stable, battle-tested, │ Vendor lock-in, recurring costs, │
│ / cloud OCR │ easy to integrate │ data residency concerns │
├───────────────┼────────────────────────────────────────┼──────────────────────────────────────────┤
│ Classical OCR │ Open-source, self-hosted, │ Requires tuning and custom pipelines │
│ (Tesseract) │ controllable costs │ for good accuracy │
└───────────────┴────────────────────────────────────────┴──────────────────────────────────────────┘