No items found.

Carlos Guzman: My Very Fast Journey From Individual Contributor to Engineering Manager

Author
Publish Date
Read Time
Datavant
September 1, 2022
I want to help build this culture.”
Photo by Natalie Fong Malis

Carlos Guzman joined Datavant in May 2022 and transitioned from software engineer to Engineering Manager within three months. Below, Carlos shares some of his experience joining the Datavant Engineering team and its high-growth, high-learning culture.

What did you do before joining Datavant?

I worked on several teams driving the direction of technology at Bloomberg in New York. I had multiple roles, but two official ones: a software engineer in an application team doing real-time pricing of derivatives (called MARS ML) to speed up valuations and delivery of risk analytics in a flexible manner for quants, and working on modeling solutions. I also worked as an Engineering Champion (approximately equivalent to staff level engineer) where I led a group of top engineers to build tools and solutions at the department or company-wide level.

How did you find your way to Datavant?

I decided to consider other options from my previous job for a combination of reasons. My career and personal growth had become stagnant, it was hard to get things done in that particular company culture, and I wanted more flexibility and autonomy. At the end of the day it’s all about trade-offs, so I was looking for the best intersection of career growth, personal growth, flexibility, compensation, and mission. I was also considering some hedge funds, which perhaps are not known to have the best company culture, but tend to compensate better. In general, this means there’s a greater incentive to work hard and possibly find greater personal and career growth. I agreed to interview at Datavant with the expectation that it might have a better culture but probably much lower pay. During the first interview and the virtual onsite at Datavant, everyone fit the values praised by the culture, something I had not seen anywhere else in any of my other experiences with other companies. The offer Datavant made was competitive (although it was hard to value the trade-offs of cash vs RSUs, options, and cash), so in the end it seemed like Datavant had the best package to offer. Before accepting the offer, I asked to talk to more people within Datavant and the pattern of people fitting the cultural values held strong. It almost felt like the trade-offs were fading away into a place where each aspect could be maximized, so I accepted the offer.

Datavant prioritizes placing new hires on real projects with real responsibility from Day 1. Tell us about your experience joining Datavant.

I joined Datavant as part of the Match Pod, a Datavant term used to describe an engineering sub-team, led by Ben Podgursky. Because of our rapid growth, that team has since split into two Pods. The objective for the Pod is to use machine learning to do PPRL (Privacy-Preserving Record Linkage), matching de-identified records for a single individual taking into account changes in records that result from changes in name, address, and phone number, or even due to typos. My first project was to set up a model inference Python service in lieu of the in-process Java implementations that historically demanded manual translation from the Python code the modelers (data scientists) wrote to Java. This project decoupled the models from Java so that no translation would be needed, which had the added benefit of being able to update the models dynamically and seamlessly.

After the initial design and some work on it, I became a manager for the De-identified Match Pod so there was less time to code directly on the project, but I wanted to offer opportunities for the team to get more involved in it for their own growth. We’re still working on the project, so “my first project” at Datavant has become my team’s project. And if that’s not a story of the dynamic growth and collaboration you get at Datavant, I don’t know what is.

“The intentionality behind this culture amazes me.”

How was this experience different from joining other companies?

Impact. In past experiences, onboarding was mainly about learning the internal tools and C++ in depth, only to join a team a couple months later and have to review again many of the concepts. This is a much slower process. At Datavant, the experience of joining the team was more similar to the experience of internships I had in the past where the learning was more hands-on. This situation also had the added benefit of a great deal of autonomy and flexibility in my schedule so that I was genuinely happy while learning, and found plenty of support wherever I looked. Personally, it was a bit hard at the beginning because I went from having all the answers at Bloomberg because of my knowledge of the internal tools to having to ask about everything at Datavant. With all this learning and a culture founded in trust, the onboarding process seamlessly merged into being comfortable with the tools and systems because I had already been working with them rather than being taught like in a classroom.

What was an early “wow” moment after joining Datavant?

My first “wow” moments were during my first couple weeks at Datavant. I kept meeting people who all fit the “smart, nice, get things done” culture. The intentionality behind this culture amazes me and I want this to be the next model for companies to aspire to. After coming back from the Seattle offsite, I told my manager I wanted to help build this culture in a managerial role in the future. (From past experience, this process could take at least a year, so I figured I’d mention it early.) To emphasize my “wowness” in this interest, I would point out that I actively did not want to be an engineering manager when I joined, but was looking to be a very senior individual contributor instead. It took two weeks for Datavant to change my mind because of the trust in others I could have. And here we are: I am now the manager of the De-identified Match Pod.

Because of the intentionality in Datavant’s hiring process, our Pod is made up of great engineers who make it easy to be a manager. Since every member of the team already has a bias for action, my two primary objectives are to remove any hurdles for the team, or to help them jump higher and further. This way, our team gets things done and we all grow together to the best of our technical, personal, and professional abilities. Since we have a diverse set of experiences, we’re constantly learning from each other while delivering the best possible product.

Carlos Guzman has a background in mathematics and computer science and joined Datavant in May 2022. Connect with Carlos via Linkedin.

Many thanks to Carlos for taking the time for this interview. Interview compiled/edited by Nicholas DeMaison.

Spotlight on AnalyticsIQ: Privacy Leadership in State De-Identification

AnalyticsIQ, a marketing data and analytics company, recently adopted Datavant’s state de-identification process to enhance the privacy of its SDOH datasets. By undergoing this privacy analysis prior to linking its data with other datasets, AnalyticsIQ has taken an extra step that could contribute to a more efficient Expert Determination (which is required when its data is linked with others in Datavant’s ecosystem).

AnalyticsIQ’s decision to adopt state de-identification standards underscores the importance of privacy in the data ecosystem. By addressing privacy challenges head-on, AnalyticsIQ and similar partners are poised to lead clinical research forward, providing datasets that are not only compliant with privacy requirements, but also ready for seamless integration into larger datasets.

"Stakeholders across the industry are seeking swift, secure access to high-quality, privacy-compliant SDOH data to drive efficiencies and improve patient outcomes,” says Christine Lee, head of health strategy and partnerships at AnalyticsIQ. 

“By collaborating with Datavant to proactively perform state de-identification and Expert Determination on our consumer dataset, we help minimize potentially time-consuming steps upfront and enable partners to leverage actionable insights when they need them most. This approach underscores our commitment to supporting healthcare innovation while upholding the highest standards of privacy and compliance."

Building Trust in Privacy-Preserving Data Ecosystems

As the regulatory landscape continues to evolve, Datavant’s state de-identification product offers an innovative tool for privacy officers and data custodians alike. By addressing both state-specific and HIPAA requirements, companies can stay ahead of regulatory demands and build trust across data partners and end-users. For life sciences organizations, this can lead to faster, more reliable access to the datasets they need to drive research and innovation while supporting high privacy standards.

As life sciences companies increasingly rely on SDOH data to drive insights, the need for privacy-preserving solutions grows. Data ecosystems like Datavant’s, which link real-world datasets while safeguarding privacy, are critical to driving innovation in healthcare. By integrating state de-identified SDOH data, life sciences can gain a more comprehensive view of patient populations, uncover social factors that impact health outcomes, and ultimately guide clinical research that improves health. 

The Power of SDOH Data with Providers and Payers to Close Gaps in Care

Both payers and providers are increasingly utilizing SDOH data to enhance care delivery and improve health equity. By incorporating SDOH data into their strategies, both groups aim to deliver more personalized care, address disparities, and better understand the social factors affecting patient outcomes.

Payers Deploy Targeted Care Using SDOH Data

Payers increasingly leverage SDOH data to meet health equity requirements and enhance care delivery:

  • Tailored Member Programs: Payers develop specialized initiatives like nutrition delivery services and transportation to and from medical appointments.
  • Identifying Care Gaps: SDOH data helps payers identify gaps in care for underserved communities, enabling strategic in-home assessments and interventions.
  • Future Risk Adjustment Models: The Centers for Medicare & Medicaid Services (CMS) plans to incorporate SDOH-related Z codes into risk adjustment models, recognizing the significance of SDOH data in assessing healthcare needs.

Payers’ consideration of SDOH underscores their commitment to improving health equity, delivering targeted care, and addressing disparities for vulnerable populations.

Example: CDPHP supports physical and mental wellbeing with non-medical assistance

Capital District Physicians’ Health Plan (CDPHP) incorporated SDOH, partnering with Papa, to combat loneliness and isolation in older adults, families, and other vulnerable populations. CDPHP aimed to address:

  • Social isolation
  • Loneliness
  • Transportation barriers
  • Gaps in care

By integrating SDOH data, CDPHP enhanced their services to deliver comprehensive care for its Medicare Advantage members.

Providers Optimize Value-Based Care Using SDOH Data

Value-based care organizations face challenges in fully understanding their patient panels. SDOH data significantly assists providers to address these challenges and improve patient care. Here are some examples of how:

  • Onboard Patients Into Care Programs: Providers use SDOH data to identify patients who require additional support and connect them with appropriate resources.
  • Stratify Patients by Risk: SDOH data combined with clinical information identifies high-risk patients, enabling targeted interventions and resource allocation.
  • Manage Transition of Care: SDOH data informs post-discharge plans, considering social factors to support smoother transitions and reduce readmissions.

By leveraging SDOH data, providers gain a more comprehensive understanding of their patient population, leading to more targeted and personalized care interventions.

While accessing SDOH data offers significant advantages, challenges can arise from:

  • Lack of Interoperability and Uniformity: Data exists in fragmented sources like electronic health records (EHRs), public health databases, social service systems, and proprietary databases. Integrating and securing data while ensuring data integrity and confidentiality can be complex, resource-intensive and risky.
  • Lag in Payer Claims Data: Payers can take weeks or months to release claims data. This delays informed decision-making, care improvement, analysis, and performance evaluation.
  • Incomplete Data Sets in Health Information Exchanges (HIEs): Not all healthcare providers or organizations participate in HIEs. This reduces the available data pool. Moreover, varying data sharing policies result in data gaps or inconsistencies.

To overcome these challenges, providers must have robust data integration strategies, standardization efforts, and access to health data ecosystems to ensure comprehensive and timely access to SDOH data.

SDOH data holds immense potential in transforming healthcare and addressing health disparities. 

With Datavant, healthcare organizations are securely accessing SDOH data, and further enhancing the efficiency of their datasets through state de-identification capabilities - empowering stakeholders across the industry to make data-driven decisions that drive care forward.

Careers

Join us on our quest to connect the world's health data

Join us

Achieve your boldest ambitions

Explore how Datavant can be your health data logistics partner.

Contact us