On October 1st I will be leaving IBM and I wanted to mark this occasion with a blog post to book-end what was a challenging, rewarding, and stimulating time in my career. It’s been a bit of a roller coaster at times, but one that has allowed me to work on a wide variety of projects and make friendships that I will hold with me long after I leave the company.
I was fortunate enough to spend my entire IBM career on research & innovation, getting to work on use cases and technologies often long before they became mainstream, such as smart assistants nearly a decade before Amazon Alexa, the ethics of data science long before a stream of whistleblowers showed us what we had been warning for years, and building a self-sovereign identity solution that was used by millions, deployed across multiple jurisdictions, supporting different credential specifications, and proving the power of self-sovereign identity at a global scale.
I joined IBM in 2001 driven by my desire to work on open source, believing in its power to amplify the impact of technology. At the time IBM was actively growing its investment in open-source technologies, and doing some great innovation in artificial intelligence, so it seemed the perfect place to park.
So let me get started and share a whistle-stop tour of my journey through IBM.
IBM LanguageWare (2001) was my first project at IBM, where I was challenged to reimagine the impact of natural language processing (NLP) and to build on decades of NLP research in IBM, to create an NLP library that could be leveraged by multiple solutions across the IBM portfolio to build an AI-powered future. LanguageWare became an integral component in dozens of IBM products – from search engines to database applications, text editors, collaboration technologies, and analytics solutions. This time in IBM was the precursor of the Jeopardy grand challenge, so a really exciting and vibrant time to be working on NLP at IBM, with lots of open source activity around the creation of Apache UIMA. During this time, I got to work with so many people who all touched me in different ways, but I have to mention the guys who started LanguageWare with me and with whom I shared many long hours working to turn our vision into reality; DJ McCloskey, Alex Nevidomsky, and Alexander Troussov.
During the decade that I worked on LanguageWare, there are too many projects to call out, however there are a couple that stand out.
1641 Depositions (2008) was the most unlikely project and yet remains in my memory as one of the most rewarding projects I’ve ever worked on. A long time before there was IBM Watson there was Father Busa, credited with being the father of digital humanities, challenging IBM founder Thomas J. Watson to build a computer program that could understand text (Roberto Busa and the invention of the machine-generated concordance). Sixty years later I was challenged by Professor Jane Ohlmeyer, Erasmus Smith’s Professor of Modern History at Trinity College Dublin, to help them understand ancient historical texts (1641 Depositions) and to piece together the facts extracted so that we could effectively uncover the truth hiding behind the story. Using a combination of our content, semantic, and network analysis, we were able to build a more complete picture of the Irish Rebellion of 1641 and help historians better understand a complex and troubled period in Irish history. Aside from being such a strong cultural and human story, it was a great challenge for our technology due to a myriad of issues in analyzing ancient text – from transcription errors to the lack of punctuation or spelling conformity in ancient English (A trip down memory lane; text-analysis jedii-style).
BBC Dynamic Web Publishing (2009) was a collaboration with the BBC to re-imagine how content gets published and build a solution that would host the 2009 Winter Olympics and eventually run the entire BBC Sports World Cup website. It was the first solution of its kind globally and served 100s millions of BBC viewers during the World Cup. It leveraged our work on graph-based semantic disambiguation from our participation in the EU Framework 6 Nepomuk Social Semantic Desktop project. This was the brainchild of John O’Donovan and Jem Rayfield from the BBC, and Martin Guillaume from IBM GBS, all of whom were just so great to work with. I have such fond memories of my time working with the guys on this project.
If I fast-forward to 2014, the next project I wanted to call-out was the Personal Social Dashboard, that combined algorithms for measuring expertise and influence with techniques for modelling knowledge flow. It was one of the most inclusive projects I’ve ever worked on, bringing together such a diverse group of stakeholders across IBM – from the employees that would use the solution to the CIO office that would deploy it, research & development that would build it, and sales & consulting that would sell it. It forced us to integrate privacy and ethics considerations at every stage of the project and gave me an invaluable education on the ethics of data science through rigorous jurisdictional privacy reviews. The solution was ultimately adopted by the IBM CIO Office, where it was used by 40,000+ IBMers. Shiri Kremer-Davidson was the key researcher out of the Haifa Research Lab that worked with me throughout this project and she was just a wonderful collaborator. Kate Ehrlich is another brilliant researcher that worked with us when we applied this technology to one of IBM’s large banking clients. They are just two of dozens of people that contributed to this project; all amazing to work with.
In 2014 I also got the opportunity to do a TED Talk on the ethics of data science, an experience that I would strongly recommend (my TED experience; so much more than I expected).
My final fast-forward is to 2020, when we were all under the shadow of the Covid pandemic and trying to figure out how to most effectively re-open economies. I was asked to apply the work I’d been doing on decentralized identity to the Covid use case, and 8 months later came IBM Digital Health Pass, a solution designed to fundamentally change how health data is exchanged; allowing individuals to hold their own data and selectively share with third parties in a way that can be trusted, verified, highly privacy-preserving, and all happening peer-to-peer and fully decentralized (A Privacy-Respectful Platform for Proving Health Status). Through my work on self-sovereign identity I got to collaborate with organizations such as the Linux Foundation’s Trust over IP and Covid Credentials Initiative and the Good Health Pass Collaborative, where I got to know so many talented and passionate people from whom I have learned so much over the years; folks such as Drummond Reed, John Jordan, Charlie Walton, Kaliya Young, Jim St. Clair, Wenjing Chu, Judith Fleenor, Andre Kudra, and too many others to mention.
Maybe it was the times that are in it and the fact that we were all locked up in our homes, working with the IBM Digital Health Pass team has been really special. It will be one of those projects that you remember in years to come with special fondness. It’s impossible to name everyone that was involved in this project, but there are a few people I personally have to call out; Eric Piscini (who was such a supportive executive throughout the project), Paco Curbera (who will go down as being one of the best managers I’ve worked for), and an amazing technical team with Rich Scott, Alan O’Neill, Corville Allen, Chandra Maduri, Gautham Velappan, and too many others to mention.
Early in 2022 I was awarded the honor of IBM Distinguished Engineer, which was a much appreciated recognition of my work at IBM over the years. Shortly after IBM decided to divest the Watson Health division to create a new company called Merative. Our US colleagues transitioned over a few months ago, and at the end of this week I, and my EU colleagues, move over to the new company. Leaving IBM is definitely bittersweet, but I will remember my time in IBM fondly and will be joining a large and vibrant community of ex-IBM friends and colleagues out in the big bad world 😉
Marie's Ramblings & Ruminations 