Hiring and retaining data science talent continues to be one of the biggest challenges facing financial services firms today. Success lies in the four Ps — People, Projects, Place and Process — and in remembering that data quality is the biggest differentiator.
- The recent AI & Data Science in Trading event in New York heard from a number of c-suite delegates on how financial services firms can build and compete for data science talent.
- The four ‘Ps’ of People, Projects, Place and Process were central to the discussions, particularly the importance of ensuring the right blend of skills across the group.
- Most data scientists move on because they are sick of waiting for data, highlighting the value of data quality in the process to build and compete for data science talent.
“Building a top-tier data science team feels like you’re trying to build a spaceship,” said Matt Granade in his opening remarks to the recent AI & Data Science in Trading event I chaired in New York.
Matt is the Chief Market Intelligence Officer at Point 72, perhaps one of the coolest jobs in the industry if you’re a data geek, like me.
One theme that kept recurring throughout the conference was the struggle to build a talented team of data professionals in financial services.
Matt stressed the evolutionary nature of recruitment and retention, suggesting that the right mix of people, process, and technology — applied to sourcing, analysis, and integration — garner the best results.
At Refinitiv, I’ve experienced the challenge of competing against the likes of Amazon, Google and Facebook for data science talent. This was also mentioned by others during the conference.
There’s only one way around this in my opinion: you need to make data the differentiator. For Refinitiv, this is a huge part of our recipe for success, and we’re now seeing the fruits of this strategy.
Sarah Hoffman, Vice President of AI & Machine Learning Research at Fidelity Investments, echoed this sentiment: “Most data scientists leave because they’re sick of waiting for data. It’s all about having a strong data strategy. Data is the foundation to the future.”
Data science talent and the 4 Ps
The 4 Ps govern the approach we’ve taken in building our Innovation Labs at Refinitiv: People, Projects, Place and Process.
First, the People are the team. It’s important to give them room to grow and showcase their work. It’s crucial to get the right blend of skills across the group — data scientists, data engineers, full-stack developers, and subject matter experts. If you have the wrong mix, you’ll completely miss the target.
Mark Antonio Awada, Chief Risk & Data Analytics Officer, Alpha Innovations, similarly expressed this: “I see lots of big firms that have a lot of data scientists, but they aren’t making money.”
Those firms, he assesses, see the quants as the “old guys” and the data scientists as the “new guys.” Mark went on to caution: “It is a mistake to marginalize the quants. Firms need to merge the data scientists with the quants, otherwise it is fatal.”
Second in the 4 Ps is Projects. It’s essential to give the team challenging and engaging projects that will benefit the company, and to allow them to use their knowledge, while also experimenting and gaining new knowledge.
Third is the Places where our Labs are located. We’re strategically placed in San Francisco, New York, London, and Singapore, with easy access to other competitive markets. We’ve created a look, feel and culture that’s compelling for our team.
We may not have exposed pipework hanging from the ceiling, but we do have great tech and a fridge stocked with tasty snacks and perhaps the odd case of beer (highest on the list of requests from the team).
Finally, it’s the Processes we employ that make us unique. It’s how we govern and prioritize our work, as well as how we engage with the various business units, and align with our internal and external tech partners.
When we stood up the new team last year, we sifted through the 60-odd projects on the list and reduced it to fewer than 10 that were hard, but important to the business. All projects have a finite length (typically 10 weeks), which means that “fail fast” is never an issue — all projects come to an end.
Some projects move on to be the basis of a production feature or business case, others are documented, and key learnings shared with the internal customer and across the business.
The integration factor
Another requirement of our operations at Refinitiv is to integrate the team with the rest of the organization. This is critical, especially as we’re not a direct revenue-generating center.
It’s important for other business leaders to see the value the data scientists provide, and the difference the team makes. This happens via a commitment to communication and close alignment with our internal stakeholders.
Related to this, Brice Rosenzweig, Global Head of Data & Innovation Group, Bank of America Merrill Lynch, spoke about the importance of a centralized data and innovation team tightly integrated into the business.
He explained how this enables people to think centrally and to augment their skills. It also allows the Lab leaders to create opportunities that span research, trading and sales, further solidifying the team’s relationships and value. Such subject matter expertise is something we value highly at Refinitiv.
Quick access to data
Joel Nathaniel Bloch, Founding Partner and Chief Risk Officer of Trinnacle Capital, discussed the huge challenges data scientists have with content quality. He shared that his team can spend 80 to 90 percent of their time normalizing and cleaning data.
Data scientists need quick access to data while the idea or opportunity is still fresh, and the business sponsor still engaged. The data scientists at Refinitiv have access to the world’s best content through an internal project that gives them an instant pathway to huge data sets, such as the firm’s machine readable news, tic data and QA Point.
We also leverage Jupyter notebooks to help light up the content quickly, as well as share code (I wish this had existed when I was a software engineer!). The data scientists receive a full working version of Intelligent Tagging, our high-performance NLP engine, and access to Knowledge Graph.
Passion and grit
While the 4 Ps get to how we run our operations, passion and grit are what we look for in talent.
Our data scientists are our organization’s trailblazers. They are exploring new frontiers. This requires a passion for that type of work and an aptitude to perform in an environment that’s full of uncertainty. Sometimes the theory just doesn’t work. They need to learn from that and move on. The outcomes will unfold.
Matt stressed similar characteristics he looks for in the data scientists he hires at Point72. He seeks people who are passionate about their work, motivated by problem solving, aspirational, and career focused; and “they have to have grit!”
Sameer Gupta, Head of Data Sourcing at Point72, underscored the significance of recruiting the right person. “I don’t believe in resumes. A 4.0 from an Ivy League institution is not a guarantee for the job,” said Sameer.
He’d rather give them a problem and throw them in the water. “If they don’t like it, they won’t come back. It’s a great way to figure out who’s right and who’s not.” This is another version of grit.
When we hire data scientists at Refinitiv, they go through a rigorous set of interviews and are given a technical challenge. It’s very much the team making the hire and as a result we have a good fit out of the box.
Data is just the beginning
The data science arena in financial services is exciting and challenging. I hope some of the points I shared will inspire your strategies for your own innovation. And, bear in mind that in the end, data is just the beginning to building a thriving team of data science talent.
We’re hiring now and have a number of exciting data science positions open. Take a look if you’re looking for a new challenge.
- AI & Machine Learning
- Artificial Intelligence
- Big Data
- Digitalisation of Wealth Management
- Elektron Data Platform
- Fixed Income Pricing Data
- Foreign Exchange Pricing Data
- Generating Insights From Data
- Guest Article
- Intelligent Tagging
- Knowledge Graph Feed
- Managing & Integrating Data
- Mergers and Acquisitions Data
- Operational Efficiency
- QA Point
- Risk and Regulatory Data
- United States