Technical Leader & Software Architect
CTO at Onlihub · Hands-on engineering
I turn business complexity into software that works in practice.
I connect business processes, system architecture and hands-on engineering — across commerce platforms, configurable business systems and AI developer tools.
Kraków, Poland
Start with the real process: people, documents and business constraints.
Define entities, relationships, rules and the boundaries between systems.
Build APIs, workers and integrations into a system that can grow.
Observe what happens, resolve bottlenecks and keep the product working.
Business context, implementation and the decisions behind each system.
View all workCommerce operations had to stay consistent across different product models, warehouses and external systems.
CTO · previously Lead Developer
The work connected the product lifecycle from catalog and inventory to orders, fulfillment and tracking, with ownership extending beyond individual services.
Read caseFrom client intake to mortgage operations
I designed and built a business operations platform used by an existing mortgage company. Conditional client forms, document OCR, staff tasks and compensation rules connected to its Encompass workflow.
The company worked in Encompass without a client-facing intake website. Staff manually gathered information, worked out which documents each borrower needed and coordinated follow-up. Employment, income and other circumstances changed the requirements, making each new application difficult to organize.
Solution design, architecture and hands-on development
The company gained a client-facing intake website connected to its internal operations. Clients could answer once and see which documents to provide; staff received assigned tasks and worked on linked records that flowed into Encompass. This simplified application handling and reduced manual coordination.
Read caseAgent orchestration, project knowledge and reusable experience
I built a service where agents find tools and project knowledge, work across connected systems and retain useful methods for later tasks.
Operational questions span several systems, while broad database access and an ever-growing tool catalog create permission and context problems for agents.
Agent infrastructure and integration architecture
I use MCP Hub to investigate errors, review communication and trace project activity across databases, logs and work tools. Relevant schemas, stored results and prior methods can be opened as needed, keeping repeated investigations focused.
Read caseFrom a business question to a working system, with clear ownership of the decisions along the way.
Translate product goals into technical decisions and an executable plan, working with stakeholders and developers while staying involved in implementation.
Read caseModel the documents, roles, calculations and workflows behind an operation, then build a system the business can use and adapt.
Read caseBuild APIs, data models and background processes that connect products and external services, with explicit rules for data ownership, caching and synchronization.
Read caseGive agents relevant project context and controlled access to working tools, with inspectable execution and reusable experience.
Read caseA few examples of turning a difficult constraint into a practical system.
CTO since November 2023; joined in February 2022 and took technical ownership as Lead Developer.
A dedicated catalog service with a Rust primary path and Python fallback. PostgreSQL retains transactional state, Redis serves repeated reads, and Elasticsearch provides the search model. Bulk processing and query optimization support this division.
Each layer has a clear responsibility.Rust / Python / PostgreSQL / Redis / ElasticsearchConditional forms determine the next question and required documents. Document fields, OCR, participant roles, tasks and status rules connect intake with staff work and Encompass exchange.
One intake starts a coordinated workflow.Forms / OCR / Workflow / EncompassProject knowledge, reusable methods and tool discovery have separate responsibilities. Exact tool schemas load when needed; permissions and execution receipts keep discovery distinct from authorization and verified results.
Selective context, inspectable execution.MCP / Retrieval / Permissions / ReceiptsStay close to the business problem and to the implementation. Make decisions understandable to the people who will build and use the system.
At Onlihub, I worked directly with the CEO on feasibility and delivery. Product ideas became architectural choices and concrete implementation work.
Read caseIn BOS, changing intake requirements became configurable form conditions, document requirements and staff tasks. Business rules stayed connected to the work they trigger.
Read caseIn MCP Hub, finding a tool, requesting approval and executing it are distinct steps. Retained results let the operator inspect what happened before deciding what follows.
Read caseI model the business, design the system, write the backend and take it through delivery.
Backend services and API design
Related workRelational models, queries and triggers
Related workCaching and temporary state
Related workSearch models and bulk indexing
Related workClear boundaries, shared business models
Related workCommerce, webhooks and synchronization
Related workAcquisition, normalization and batch processing
Related workTool discovery, context and orchestration
Related workFrom documents to configurable business rules
Related workTechnical depth, business understanding, and hands-on execution.
Catalogs, marketplaces, warehouses, orders and fulfilment: the shared model and the boundaries between them.
Designed shared product, inventory and order models for Onlihub and MySellerHub, connecting channels with different warehouse models and external identifiers.
Follow the product references from a channel listing to warehouse assignment.
Keep the product model distinct from its listings and external identifiers.
Choose a node to inspect the decision behind it.
Follow the product references from a channel listing to warehouse assignment.
MySellerHub: a stock update can finish after the desired quantity has already changed.
Connect products, SKUs, variants and listings while preserving each channel’s identifiers and rules.
Connect stock, reservations, customer and supplier orders, shipping and tracking across warehouses.
Handle repeated webhooks, rate limits and partial failures with retries, idempotency and reconciliation.
For a mortgage company, I connected client intake to documents, OCR, staff tasks, compensation and Encompass. Configurable forms and business rules let the same core support different operating scenarios.
A client’s answers select the next questions and document requirements. Uploaded files supply field values through OCR; document events assign participants and tasks. Staff work from personal Kanban boards, while records and selected statuses flow to Encompass. Calculations and compensation retain their source relationships.
BOS: document events select concrete rules for participants, tasks, forms and integrations.
Typed fields, relationships and form answers give business data an explicit structure.
Choose a node to inspect the decision behind it.
BOS: document events select concrete rules for participants, tasks, forms and integrations.
Calculation is a separate route, with its own inputs, preview and result document.
Define document types, fields and parent–child relationships. Connect organizations, departments and memberships to roles and access conditions.
Collect answers in multi-step forms and preview formula results. Document events run configured participant, task, form and integration rules; calculation flows create result documents.
Link internal accruals and deductions to a document, recipient and rule. Keep calculation sources and integration-run records so a result can be traced back to the operation behind it.
Large-scale parsing with concurrency and proxy pools; Elasticsearch bulk reads and writes; separate database, cache, read and worker paths.
Amazon data: parallel collection and Elasticsearch batch operations. Catalog: Rust reads, Python fallback and Redis caching. Onlihub: business modules with explicit application contracts, repository ports and infrastructure adapters.
Amazon data work: parallel collection, grouped storage operations and search-oriented data.
Choose a data path for the workload: collection and persistence, or low-latency reads.
Choose a node to inspect the decision behind it.
Amazon data work: parallel collection, grouped storage operations and search-oriented data.
Catalog: cached data, a fast Rust read path and a Python fallback.
Coordinate asyncio, processes and threads, account/proxy pools, throttling and retries for large catalogs.
Normalize and deduplicate data, then batch Elasticsearch indexing and reads to avoid thousands of small storage calls.
Separate transactional data, search/read paths, Redis cache and background workers; diagnose bottlenecks with query plans and logs.
MCP Hub combines tool discovery, local embeddings and reusable methods with controlled execution across databases, logs, tasks and communication.
Implemented hybrid knowledge retrieval, optional local reranking of methods and experience feedback. On-demand schemas, retained results, shared runner budgets and persistent automations keep execution and context manageable.
MCP Hub: find context and the right tool, then execute with a verifiable result.
Start from a concrete project task and the scope of access it requires.
Choose a node to inspect the decision behind it.
MCP Hub: find context and the right tool, then execute with a verifiable result.
After completion: submit a useful method for retrieval by a later task.
Find tools, knowledge fragments and relevant methods; open exact schemas and larger results as needed.
Define permission boundaries, approval points and repeat-operation behaviour before tools change real systems.
Check results against execution receipts; retain useful methods and feedback with their applicability and observed outcomes.
Start with a business goal, work out a solution and carry it into code. My work combines architecture, explaining alternatives to stakeholders, hands-on development and coordination where the project needs it.
For BOS, I proposed and developed the solution, explained its architecture to the business and refined it around real workflows. At Onlihub, I took over architecture, database, backend and integrations after the lead developer left, coordinated development and advised the CEO on feasibility.
A technical decision connects a business constraint to a concrete implementation.
Start with the business process, its constraints and who is responsible for the outcome.
Choose a node to inspect the decision behind it.
A technical decision connects a business constraint to a concrete implementation.
Observe a bottleneck, make a focused change and review the effect.
Work with the business to understand the process, its constraints and what a useful result would change.
Translate the model into database, backend and integration decisions; explain the choices while implementing them.
Coordinate tasks and development, handle production problems and keep improving the product after launch.
Manufacturing, CAD/CAM and CNC built the foundation: understand materials, operations and constraints before automating them.
Prepared parts, machine programs and technical documentation for serial production; refined operation sequences and material use, then automated production planning and accounting in 1C.
CAD, machine programs and documentation connect a design to a manufactured part.
The part connects design intent to materials, operations and manufacturing constraints.
Choose a node to inspect the decision behind it.
CAD, machine programs and documentation connect a design to a manufactured part.
The same physical process becomes structured materials, plans and accounting data.
Model parts and assemblies, choose manufacturing operations and account for materials and tooling.
Prepare machine programs, postprocessors and technical documentation for cutting, turning, milling and punching.
Refine operation sequences, material consumption and bottlenecks as products move into serial production.
Independent product work across game backends, TON, Unity physics, modular vehicles, procedural worlds and 3D assets.
Led TON Tanks product development and built its backend and game logic. Independent Unity work explores modular tanks, tracked physics, damage and procedural city generation.
A turn-based server battle connects player actions, rules and client updates.
Two independent examples: a server-side battle and a modular Unity vehicle.
Choose a node to inspect the decision behind it.
A turn-based server battle connects player actions, rules and client updates.
A separate Unity project explores modular construction, tracked physics and a 3D world.
Connect game rules, progression, rewards and ownership with the backend and product concept.
Explore modular vehicles, damage, tracked physics, multiplayer and procedural environments in Unity.
Create and optimize models, materials, textures and engine audio banks for use in a real-time game.
Architecture, technical leadership, or a product that needs a solid foundation.
Telegram Mini AppThe portfolio, inside TelegramBrowse projects and send an offer in the Mini App. Open the bot, then choose the app.Open the botA role, a product, or a technical challenge. Share a little context.
To use the form, enable JavaScript. You can also write to me directly on Telegram.
Write on Telegram