Sign In

Sofus Macskássy

Data Science Leader, Advisor and Mentor - Data Science @ LinkedIn

Professional Background

Sofus Macskássy is a highly accomplished data science executive renowned for his exceptional leadership and expertise in guiding teams to harness the power of data. With extensive experience directing scientific inquiry and spearheading innovative product development, Sofus has established himself as a visionary in the field of data science. His career spans numerous roles in major tech firms, including his current position as the Head of Data Science Research and Productivity at LinkedIn. In this role, he leverages his deep understanding of user interactions and digital engagement to enhance company performance and drive innovative solutions.

Throughout his career, Sofus has showcased a remarkable ability to build and lead data teams, demonstrating his knack for creating productive environments where creativity and analytical skills can flourish. He has a particular interest in user modeling, information filtering, managing online identities, and how individuals derive value from their digital interactions. This unique perspective informs his strategic approach to data science, making him a sought-after expert and advisor for organizations looking to capitalize on big data.

Sofus's career is marked by significant contributions to both private sector endeavors and academia. He has an unwavering commitment to fostering the next generation of data scientists and frequently shares his knowledge and insights through mentorship roles, including his past advisory position at Alchemist Accelerator. His collaborative spirit and dedication to education were further exemplified through his past involvement as an Advisory Board Member for the MS Analytics program at Georgia Institute of Technology.

Education and Achievements

Sofus Macskássy earned his foundational knowledge in data science through rigorous academic training at Rutgers University, where he obtained his Ph.D. and Master's degrees, along with a Bachelor's degree from Rutgers College. His academic achievements set the stage for a career characterized by a blend of theoretical knowledge and practical application in the real world.

In addition to his academic credentials, Sofus has authored and presented numerous papers during his tenure as an Assistant Research Professor at the University of Southern California and as a Research Scientist at New York University. His efforts in academia contribute to a broader understanding of data science applications and are a testament to his commitment to the field.

His leadership roles extend to important organizations in the tech industry, where he has made lasting impacts. As the VP of Data Science at HackerRank, Sofus was instrumental in leveraging data-driven insights to enhance technical recruitment processes. His previous roles include Head of Data and Analytics at Branch Metrics and Manager of Applied Machine Learning at Facebook, where he honed his expertise in managing core data science operations.

Sofus's contributions to the field have not gone unnoticed; he served as the General Chair for SIGKDD 2014, a pinnacle conference in knowledge discovery and data mining, reinforcing his status as a leader in data science.

Notable Engagements

Sofus Macskássy's career includes a multitude of distinguished positions that demonstrate his versatile skill set and leadership capabilities. Not only has he built teams from the ground up, but he has also successfully scaled existing teams to achieve even greater heights. His experience working with startups, academic institutions, and industry giants gives him a unique perspective on the challenges and opportunities within data science.

In his role as Limited Partner at Essence Venture Capital, Sofus contributed his valuable insights to emerging tech companies, emphasizing the significance of data-informed decision-making processes. Engaging with these startups has further broadened his understanding of the data landscape and the driving factors behind successful tech innovations.

As an Adjunct Professor at the University of Southern California, Sofus has also taken strides in academic leadership, teaching the next generation of data scientists and inspiring them with real-world applications of data science principles. His commitment to lifelong learning and education is evident in his ongoing involvement with various educational initiatives and mentorship programs.

Throughout his various roles, Sofus has remained energetic and passionate about the transformative power of data. He has cultivated a reputation as a go-to expert for organizations eager to understand the complexities of user behavior in digital landscapes. His insight into how individuals interact with technology not only benefits his team at LinkedIn but also resonates across companies seeking to leverage data for competitive advantage.

Conclusion

In summary, Sofus Macskássy is a prolific data science executive whose career is characterized by leadership, innova

Highlights

Apr 3 · linkedin

Where does data curation fall into your data strategy? And what, in your mind, even is data curation?

I am glad to see an increase in discussions on the importance of data quality and what data is available to our LLM- and AI-driven products during (pre)training as well as at inference time. However, I am also starting to see everybody wanting to be part of this discussion in any way they can which I think is diluting our overall understanding of this important topic.

Here are my thoughts on what big components are needed in your Data Curation strategy:

  • Metadata: You need metadata on your data to know what you have. This includes flagging and identifying sensitive information that needs special handling for compliance and data governance.
  • Cleaning: You need to clean your data to make it valuable. Remove outliers, duplicates, normalize data, merging items for ground truth (e.g., entity resolution).
  • Knowledge graph: Extract knowledge from your datasets, build relations between them, and codify them into a semantic knowledge graph to get much deeper insights into your data.
  • Indexing: You need to index your data to support data discovery. This is a big and complex area in itself, but includes chunking pieces of text extracting phrases and text, finding key pieces of information with appropriate context and indexing them in whatever format makes sense (e.g., traditional search infra, key/value store, vertical DB, graph DB)
  • Monitoring and alerting: A lot of the data that should be curated comes from streaming data that comes from many sources including your own data pipelines (including your curation pipelines), ingestion of third-party data and customer interactions. The dynamic nature of this data requires you to monitor this data and alert you whenever something significant is changing so you can address it as needed.

Once you have a good data curation strategy and are executing on it, then building AI products, performing opportunity analyses, getting insights for decision-making, and much much more becomes a whole lot easier, robust and trustworthy.

Am I missing anything? Would love to hear your thoughts.

#dataliteracy #datacuration #dataquality #datacleanliness #datadiscovery #compliance

Mar 21 · linkedin

At LinkedIn, we’re committed to doing our part to make the global workforce a more equitable place where skills are taking the primary seat over pedigree and your personal network. Our Knowledge Graph team is doing this by building skills-first experiences directly in our products via our Skills Graph.

In order for both sides of the marketplace to get the benefit of skills, we need to speak the same language. At LinkedIn we are building the Skills Graph that does exactly that. Super happy to see this blog from our taxonomy team that speaks to how we build and maintain this taxonomy.

Thank you to the skills-graph team for sharing this excellent piece of work.

Yi Pan, Shuang Carol J., Shiyong Lin, Xiaomin W., Michael O'Neill

#skillsfirst #skillsbasedhiring

Related Questions

How did Sofus Macskássy develop his expertise in data science and team leadership?
What are the primary contributions Sofus Macskássy has made to online user modeling and interaction?
In what ways has Sofus Macskássy impacted product innovation at LinkedIn?
What insights did Sofus Macskássy gain from his experiences as a Limited Partner at Essence Venture Capital?
How does Sofus Macskássy continue to influence academia through his involvement with educational institutions?
Sofus Macskássy
Add to my network

Location

Menlo Park, California