Mac – React, Python, Django

Tele health platform for US market. HIPPA compliant.

• Built a HIPAA-compliant telehealth platform focused on women’s metabolic and hormone health from the ground up.
• Designed and integrated a production-ready infrastructure stack across Next.js, Stripe, AWS, Prisma, and healthcare workflows.
• Turned a complex healthcare idea into a real operational product with subscriptions, onboarding, and patient flows.
• Combined engineering, AI, compliance, SEO, and product strategy into a single fast-moving execution pipeline.

Developing and evaluating chatbots/agents for Essay Scoring for a University of Toronto accountancy courses. Transparent, explainable, and bias-aware AI assistant for education, healthcare, and media, with a focus on trustworthy human-centric artificial intelligence.

• created agent-based workflows using discriminator-based approach
• created evaluation harness for assistants agents - Ragas, DeepEval, LangSmith with the help of human annotated data sets - Gold Standard
• combining Machine Learning approach with NLP approach for agentic workflow creation - use of PyCaret, Zeno, WeightsAndBiases, MlFlow
• detecting and handling bias

B2B AI-powered SaaS bidding platform for film industry in the USA. The service allows for automated bid creation for film producers based on the data sources provided.

• Was the first out of total of five developers on the project;
• Architected the foundations for back end, front end, and AI components of the app;
• Explored and integrated the latest NLP tools, MindsDB, HuggingFace open-weight models and OpenAI API,
• Created and maintained DevOps pipelines;
• Backend development from scratch (MongoDb, Nest, Atlas) data modelling, query optimisation, indexes, aggregations, DDD Schema design, migrations
• Cooperated with the product stakeholder to maximize the value delivery with given resources.

Creating AI-first EHR for a private hospital in Poland. Automating document creation and management with the help of speech2text and text2speech models.

• creating AI-first EHR for document creation, using both open-weight and close-weight speech2text models, Wisper model.
• Agentic AI used for creation and management of the document workflows, RAG, PineCone, LangChain, LangGraph
• Adding a solid evaluation framework around the use of AI based on the human-annotated data, DeepEval
• Combining rule-based approach with LLM approach for decision making,
• Strictly applying compliance rules for the medical sector tech

A receipt scanner app with 200k+ downloads (Android+Apple).

• Provided cost-effective solutions for a marketing campaign using a variety of tools: no-code, low-code, and in-house. Only recommending in-house solutions when it was viable from the business point of view to do so;
• Designed and implemented an email marketing campaign;
• Created ETL data pipeline for fetching and cleansing the data for the campaign;
• Provided landing page upgrade (React) to match SEO goals;
• Integrated tracking tools on all stages of the campaign funnel (Amplitude);
• Improved the CTR on the landing pages through keyword analysis and design changes;
• Researched and implemented PoC for go-to-market strategy - integration of of hunter.io / apolo.io / zoominfo sales OS data sources (Python/Javascript).

Educational platform for aspiring chefs in French, Italian, and Indian cuisines. Heavy use of stable diffusion (ControlNet) and LLMs tools (LangChain) for content creation.

• Created front-end and back-end architecture;
• Successfully integrated with LangChain to generate content;
• Incorporated Stable Diffusion for text-to-image editing capabilities;
• Integrated message brokers to enable efficient asynchronous processing.

The company's mission was to craft a learning app tailored for individuals on the autistic spectrum. In pursuit of this vision, the project had a dual focus:

• To establish seamless integration with education providers' APIs, enriching the learning experience.
• To create an environment free from distractions, extend this commitment beyond the app to encompass the operating system itself.

• Laid the foundations for the app Front-end and Back-end;
• Created Django Rest Framework API;
• Designed the database schema;
• Created Front-end component-driven architecture with Sass and React.

AI powered virtual assistant, chatbot for a real estate platoform in the UK

• pre-LLM (pre-transformer) natural language processing, creation of chatbots
• combining rule-based approach with NLU, integrating RASA framework
• automating customer support using NLP tools: NLTK, Dialogflow, RASA
• integrating chatbots with email (mailgun), sms (Twilio) and whatsApp
• use of knowledge bases (KB), named entity recognition (NER) and intent classification
• created a recommender system for real estates to be viewed by prospecting clients

White label crowdfunding platform in the UK. Their financial activities are rigorously overseen and compliant with the stringent regulations set forth by the British Financial Conduct Authority (FCA). With a substantial clientele comprising numerous banking institutions, their utmost priority lies not only in meeting regulatory requirements but also in ensuring the highest levels of security for their products and services.

• Successfully transitioned the product from its MVP stage to a fa full-blown commercial product;
• Demonstrated expertise in full-stack development, spanning from data migration to HTML and CSS;
• Effectively reduced maintenance costs by implementing a multi-tenant architecture;
• Implemented Test-Driven Development (TDD) practices within the product development cycle, including the addition of unit, integration, and end-to-end test setups;
• Pioneered the integration of DevOps practices into product development cycles, including the creation of the CI/CD pipeline and deployment of small chunks with rollback capabilities;
• Designed and created a web API.

The biggest EPoS supplier in the UK. Provision of order-, stock- and staff-management services available on customer premises. Integration with payment providers and banks.

• Managed the synchronization of data across tens of thousands of EPoS devices for each customer, C#, MSSQL, Azure;
• Created a data merger responsible for two-way synchronization of the Head Office with the site offices;
• Created technical foundations for loyalty program for pubs and restaurants;
• Created kitchen management software for pub and restaurants involving: Azure Cloud, MongoDb, AspNet, C#;
• Created a ticketing system in Django, Python.