Howard Melnyczuk

Full-stack Developer

b. 1989, British

I am a full-stack developer and creative technologist with proven experience working on large-scale established systems and brand-new greenfield projects. I am equally at home in large multi-team companies as I am in small dynamic agency environments, whether as a freelance contractor or an integral team member. I'm now focused on tech-for-good, fighting climate change, and building a better future.

Contact

  • h.melnyczuk@gmail.com
  • github.com/melnyczuk

Education

  • 09.2016–09.2017
  • MA Computational Arts
  • Goldsmiths, University of London
  • 09.2009–06.2012
  • BA Cinema & Photography
  • University of Leeds

Selected Relevant Experience

Human Health

01.2023–08.2023

Frontend Developer Typescript

As part of a short contract, I aided the core development team in bringing an MVP healthcare application to market during the prelaunch crunch. As a product designed to provide further communication between patients and clinical professionals, the implementation of best practices around data security and user privacy was a top priority. Joining a globally distributed team, this role required excellent asynchronous communication, using agile methodologies to manage my tasks and time effectively.

  • A cross-platform React application bundled for web and mobile distribution.
  • Data security & user privacy best practices
  • Best practices for accessibility and internationalisation (EN/ES)
  • Agile & asynchronous working environment

Aware Health GmbH

08.2021–01.2023

Lead Backend Developer Typescript, Python

As the first engineer on the project, I collaborated closely with a design agency for the initial phase of research & development. This required developing the first iteration of the mobile app using React Native, whilst architecting and building MVP backend services to support functionality. Following this initial contract, I then became a full-time team member to aid the growth of the in-house engineering team.

I researched, prototyped and built a backend ecosystem of hexagonal micro-services using a domain-specific architecture that was intuitive to develop, iterate and test. As a healthcare product, data security and privacy were paramount.

These services communicate using Redis PubSub to provide an asynchronous worker/broker queue that is fast, scalable and secure within the private network. Each service is containerised for easy deployment to the cloud. These services include:

  • A Node service written fully in TypeScript with Postgres, wrapped by a thin Express REST API layer
  • A data-processing Python service that uses OCR & OpenCV to extract health data from user-generated photographs of printed medical documents
  • A Node CMS to enrich user content and support the needs of a growing in-house content team
  • An adapter layer service to connect with partnering medical labs

TransferWise

10.2019–08.2020

Full-stack Developer Typescript, Java

As a full-stack developer in the Receive team at TransferWise, I worked to improve the user experience when receiving international transfers and setting up direct debits. I worked to make complex user journeys into smooth seamless experiences using React and Angular, and I worked on system-critical Java Spring Boot microservices that adhere to the standards of financial authorities worldwide.

  • Built & maintained Java micro-services using the Spring framework
  • Developed Typescript React apps and contributed to front-end guild tooling
  • Collaborate in product discussions & input on product decisions
  • Continuous deployment and inspection on AWS

BBC

09.2018–08.2019

News Labs - Full-stack Developer Typescript, Python

News Labs is a technology-agnostic, project-driven innovation incubator based across the BBC's R&D and News departments, focusing on cutting-edge projects addressing concerns facing modern journalism. During my time on the team, I developed a News Discovery tool that clustered user feedback to provide topic insights, and I implemented a package intended to improve analytics from user-facing prototypes. I also organized ways of working sessions to promote agile methodologies and attended training on security, threat modelling & test-driven design.

  • Rapid prototyping in React and Python
  • Project ideation for greenfield projects
  • Led the development of an ML/NLP tool for news discovery
  • Architected projects for deployment on cloud infrastructure

iPlayer - Frontend Developer Typescript

I worked to improve the multitude of React apps that work to form the BBC iPlayer website, one of the UK's most used services (with up to 21 million requests an hour). As part of a cross-functional agile team, I worked closely with UX designers to implement a rebuild that ensured universal accessibility, Testers to ensure full browser support across all operating systems, and the product owner to design insightful A/B tests.

  • Built and maintained multiple isomorphic React applications
  • Developed sophisticated A/B tests with insightful metrics
  • Worked with UX designers to implement new features
  • Commitment to accessibility & cross-browser support
  • Updated REST client to improve caching

Selected Personal Research Projects

2021   Wool-gather   TypeScript, Python

Wool-gather was a project researching personal responses to machine intelligence and non-human creativity. The work exists as a standalone executable, consisting of a Typescript Electron app with a self-contained Python server spawned as a child process. The application feeds frames from a real-time webcam feed to an object detection model, the output of which is then used to generate meandering folk tales using a custom-trained large language model (LLM). The outputted text was then composited with the live feed as part of a site-specific installation.

2020   Supercollager   Python

Supercollager is an ongoing project exploring photography and image-making in dialogue with machine learning and computer vision. It is built in Python as a native executable CLI and as a Flask microservice. It uses OpenCV, relying on the MaskRNN image segmentation model.

2018   Object Permanence   JavaScript, Node.js

Produced during a residency at the Victoria & Albert Museum, Object Permanence is an investigation into the legacies of colonial atrocity that exist in institutional archives. By cross-referencing historical dates, names, and places against key fields in the V&A’s digital database, the work draws out objects from the museum's collection acquired during acts of British colonial aggression.

2018   Gail (commission for the Bureau of Meteroanxiety)   JavaScript, Python

Gail was a chatbot developed as part of The Bureau of Meteroanxiety, an immersive theatre experience for Next Wave festival, Melbourne in 2018. It comprised a React client and a Flask app written in Python that used Natural Language Processing and Machine Learning to automate real-time responses to user questions.

2017   Rrosetta   Python

Rrosetta was a Django application using natural language processing and web scraping to generate personalised zines. With the consent of the participant, the application was able to retrieve their sent emails using the Google Mail API. Drawing self-publishing into wider debates about privacy, automation, and AI, this project set out to examine questions about the price, cost, and value of identity during a time when data privacy is a critical concern.

Skills

Backend

Django, Flask, Express, Spring, Prisma, Postgres, MySQL, MongoDB, REST, GraphQL (inc. Apollo), Websockets, Jest, Gunicorn, Redis (inc. PubSub), Celery, SQLAlchemy, NumPy, OpenCV

Frontend

React, React Native, Next.js, SwiftUI, JetPack Compose, Redux, Webpack, D3, Three.js, Scss, Less, Emotion, Enzyme, React Testing Library, Netlify

Best Practices

Git, UNIX, Docker, FP & OOP, Agile & Scrum, AWS, GCE, OWASP & threat modelling, Testing with TDD & BDD