I’m Mark Tsimelzon, and this is my first blog post as Syapse’s VP, Engineering. I joined Syapse three weeks ago from Yahoo, where I led a sizable team building petabyte-scale data systems. In this post, I’d like to tell you why I decided to join Syapse.
I will not write about how meaningful it is to work for a company that builds software for doctors who treat cancer and other genetic diseases, even though this is a huge part of everybody’s motivation at Syapse. Nor will I write about the promise of precision medicine—there are plenty of other people at Syapse who are experts in this field. I am not.
Instead, I want to write about the exciting technical problems that we are solving at Syapse, and how these problems are both similar and different from the challenges I worked on in the startups and large companies I previously worked at. There are three areas that I would like to cover: scale, complexity, and speed of innovation.
Let’s start with scale—an area where I have considerable experience. At Yahoo, Akamai, and Coral8 (the startup I founded), I built big data and real-time systems for financial institutions and for very large web sites. I led teams that built core Hadoop functionality and leveraged Hadoop for petabyte-scale applications.
In healthcare, the data scale is currently more modest, but this is where things start to change with molecular testing. Given the exponentially dropping cost of testing, the rapidly increasing number of genomes tested, the longitudinal nature of these tests, and the fact that a human genome contains 3 million variations from the reference genome, the amount of patient data hospitals must manage and query is increasing by orders of magnitude.
The lesson we have learned from building data platforms in many domains applies here: to handle new kinds of data, new data scales, or to handle real-time data, new systems need to be introduced. Older platforms cannot be easily adapted to satisfy the new requirements. For example, the business intelligence industry learned that traditional operational data stores cannot support data warehouse workloads. Syapse’s opportunity in healthcare is very similar.
Now let’s turn our attention from scale to complexity. This is the area where Syapse’s challenges are so much greater than those of any web company out there. Most web companies’ data are quite simple: they primarily record click-streams and view-streams. They care about what their users see, and what they click or buy. (It’s a bit more complex than this, but not by much.) This is why web companies store their data in relational or even NoSQL data stores.
Syapse, on the other hand, deals with medical data and genomic data. Human biology and medicine are vastly more complex fields. Information about genes, patients, tests, diseases, treatments, clinical trials, and so forth cannot be easily mapped into a relational model, let alone NoSQL models. This complexity is why Syapse turned to the Semantic Web stack, and why we store data in a highly optimized RDF store. We use SPARQL, not SQL, to query RDF data. This allows us to build much more interesting and sophisticated applications, even in the presence of stringent regulations. For the first time in my career I feel I’m working not just with data but with knowledge.
And finally, let’s turn to the most exciting area: the speed of innovation. This is where Syapse wins hands down. The industry’s understanding of basic web or e-commerce sites is pretty mature. It is unlikely that any new technology will improve business metrics associated with web or e-commerce by a huge margin. In a mature field, often even changing some KPI by 1% is considered a big achievement.
Precision medicine, however, is a rapidly developing field. We learn more about the human genome every day. New study results are published all the time. New tests, clinical trials and therapies are constantly introduced and repurposed. Therefore, it’s imperative that Syapse builds a platform that can sustain this speed of innovation. The Semantic Web stack I already mentioned helps a great deal, and so does our SaaS model. We can integrate new knowledge into our application, and we can also quickly and securely integrate with hospital EMR systems without deploying any software on their premises.
I look forward to helping Syapse address these technical challenges, bringing my 20 years of experience in web, e-commerce, and finance to bear. These challenges are incredibly exciting for any technologist. Combined with working on something as meaningful as treating cancer patients more effectively and the amazing Syapse team, one can hopefully see why joining Syapse was one of the easiest decisions of my career.