My tech
stack.

My primary language for backend, automation scripts and AI projects. I use it with Django, FastAPI or Flask depending on project scope — from full management systems to data-processing pipelines.

Framework of choice for systems with structured data, authentication and back-office admin panels. Django's ORM, migrations and auth system eliminate weeks of boilerplate — ideal for enterprise management platforms.

High-performance REST APIs with strict typing via Pydantic and automatic Swagger docs. I reach for FastAPI when frontend and backend need an explicit API contract — especially for apps with maps, dashboards or future mobile clients.

Ideal microframework for modular systems built with Blueprints. When the problem domain is complex and I need full control over the architecture, Flask gives me the flexibility Django doesn't always allow.

For complex interfaces with heavy dynamic state — interactive dashboards, multi-step flows and SPAs with real-time updates. Part of the stack for projects currently in development.

Vanilla JS for most web projects: faster to load, no dependencies to maintain and powerful enough for 90% of use cases. I combine it with native browser APIs (Fetch, IntersectionObserver, localStorage).

PostgreSQL in production for systems requiring transactional integrity and complex queries. SQLite in development and for moderate-scale internal systems (WAL mode resolves most concurrency cases without Postgres complexity).

Shopify with Liquid for e-commerce stores that need customization beyond stock themes. Line Item Properties for custom variants, dynamic sections and checkout flows tailored to each business.

Primary tool for interface design, flow prototyping and design system building. I use Figma in the planning phase of every web project before writing code — wireframes and prototypes reduce costly changes during development.

Affinity Designer, Photo and Publisher for graphic design projects requiring professional print output: brand manuals, stationery and editorial material. A professional Adobe alternative for projects that don't need the full suite.

Adobe suite for brand identity, editorial design and visual content. Illustrator for logos and vector systems, Photoshop for retouching and composition, InDesign for stationery and brand guides, After Effects for motion graphics.

Groq delivers language model inference up to 10× faster than other APIs, with tokens-per-second that make real-time AI responses possible. I use it for conversational assistants and text analysis in production.

Google's API for Gemini models — especially useful for multimodal use cases (text + image) and when integration with the Google Cloud ecosystem is required. Part of the stack for projects currently in development.

Meta's open-source language model. Its main advantage: it can run locally or via Groq without proprietary per-token costs. Ideal for conversational assistants in management systems where latency and cost are critical.