I'm Mike Lynch, a developer and analyst who works on specialised IT support for academic research, research data management, open access data and data publication. I'm research engineer group lead at the Sydney Informatics Hub, a team of data scientists, statisticians and software engineers who provide specialised support to researchers across all disciplines at the University of Sydney. You can get an idea of the sorts of things I'm interested in from my most recent presentation, Real Programmers Don't Write Research Software
A more formal version of my CV as a PDF
Open research data
I'm an active participant in the development of RO-Crate, an open standard for storing rich, machine- and human-readable contextual descriptions for research datasets, and have contributed to the node.js libraries which support it.
I'm also a participant in the community supporting the Oxford Common File Layout, OCFL, an open standard for data-centric, sustainable digital repositories, and I'm the lead developer of ocfl-js, a node.js library for working with OCFL repositories.
I'm the lead developer of Oni, a simple toolkit for building faceted search interfaces to OCFL repositories. Oni is in production as the UTS Data Publications Portal and has also been used for specialised research data collections.
I'm also a contributor to Arkisto, a framework for sustainable research data repositories with participants in the international research community which is based on OCFL and RO-Crates.
- RO-Crate: GitHub, npm
- OCFL: GitHub, npm
- Oni: GitHub
- Arkisto: an open-source, standards based framework for digital preservation - my presentation at Linux.conf.au's GoGLAM mini-conference in January 2021
- FAIR Simple Scalable Static Research Data Repository Demonstrator - a poster presentation on an early version of Oni from eResearch Australasia 2019
Research data management
I have worked in the eResearch Support Group as an analyst and developer for the past ten years, and helped to establish UTS' research data management catalogue with two grants from ANDS in 2011.
In this version of ReDBox, we introduced the key new feature of workspace provisioning, which allows projects in research platforms such as LabArchives, GitLab, REDCap, OMERO and others to be provisioned via API and linked to research data management plans.
I'm currently the UTS lead for our participation in the ARDC's Institutional Underpinnings program, a collaborative effort to coordinate research data management best practices with participation from twenty-five Australian universities.
- ReDBox research data management planning tool
- Provisioning and Sustaining Research Workspaces and Repositories - slides from my presentation on ReDBox 2.0, delivered at eResearch NZ in Auckland in 2019
- Institutional Underpinnings
Research sector knowledge
In 2010 I was the technical lead for the UTS ERA team, and developed systems to enable timely delivery of reports on the shape of our submission to academic leaders. Our team was awarded the Research Support award in the Vice-Chancellor's Awards for Research Excellence awards in 2010 for our efforts on the ERA submission. I have extensive knowledge and experience of the Australian tertiary sector, having supported research at the faculty, research office and IT department level.
I have used my programming skills both to develop and maintain the eResearch group's applications and infrastructure, and to help researchers solve their technical problems at our weekly IT drop-in sessions.
I have hobbyist experience with Haskell, Go, Lua and sclang.
Program management and systems architecture
Our team was one of the first in ITD to shift to an agile development methodology and we've been developing our research data management and publication platforms using agile for the past three years.
In my work for a research support team within a IT department, I have experience of delivering support for specialised research software systems and toolchains, while still ensuring that systems meet the standards expected in modern IT practice. Our team has pioneered the use of specialised container platforms such as Singularity to standardise and streamline the deployment of research software.
DevOps and systems administration
My team at UTS has pioneered the use of devops technologies such as Docker and Ansible to deploy applications both to our on-premise compute and to the AWS cloud. I have experience in developing options papers for the deployment of on-premise applications to different cloud options. I have Linux systems administration skills and have used our high-performance computing cluster, giving me exposure to GPU programming using the CUDA platform and job scheduling with PBS Pro.
Presentation, training and management
I have presented at national and international conferences on research data management, metadata standards, and research support project management. As part of my work on ERA I presented regularly to researchers and academic leaders. I am a regular member of the UTS Researcher Development Steering Committee and the ITD Solution Design Centre.
I participated in ResBaz Sydney in 2017, an annual event promoting digital skills for researchers, and am one of the UTS coordinators for our participation in the online ResBaz Sydney 2021.
I am a regular attendee and presenter at eResearch Australasia and have also participated in eResearch New Zealand and Open Repositories.
I've acted as manager for teams in the Information Technology Department and in the Research Office, which has given me experience of adminstration and the university budget cycle.
Natural language processing
I have used the Python natural language toolkit both as part of my role in supporting and training researchers, and for my own creative projects. I have developed and presented a Text Mining for the Humanities workshop which demonstrates the basics of text tokenising, cleaning and categorising using NLTK and Google CoLab. As a creative outlet, I've participated in the NaNoGenMo (national novel generating month) event, using a range of different programming languages and techniques, from machine learning to a combinator library written in Haskell to generate the history of a fictional empire.
- Text Mining for the Humanities
- Excavate - a modification of J C Johnson's torch-rnn library which allows the unearthing of a hidden narrative from one text using an RNN trained on another text
- TextGen - a Haskell combinator library for text generation
- NaNoGenMo - my entries for the annual tongue-in-cheek novel generation contest
- Twitter and Mastodon bots
I have used several machine learning platforms as part of creative projects. Neuralgae is a generated graphic novel based around images generated using the Caffe framework. Glossatory is a neural network trained on the WordNet lexicon which generates absurd dictionary definitions, running on J C Johnson's torch-rnn, a Lua-based recurrent neural net. I have also had experience with the TensorFlow machine learning platform.
Music and digital signals processing
As an amateur musician, I've taught myself synthesis and signal processing with the SuperCollider platform, which has given me considerable experience with its sclang language.
2015-present: eResearch support analyst, University of Technology Sydney
2014-2015: acting manager, eResearch Support Group, University of Technoloy Sydney
2010-2014: eResearch support analyst, eResearch Support Group
2009-2010: acting manage, Research Performance and Monitoring, Research and Innovation Office, UTS
2003-2009: web developer, Faculty of Engineering and Information Technology, UTS
2001-2003: freelance web development, home duties
References available on request.