In the healthcare sector, digital innovations are multiplying, but embedding them into the operational reality of professionals remains a challenge. Robin Sarfati, co-founder and CTO of Tekkare, embodies the drive to create technological solutions that are simple, reliable, and tailored to real-world needs.

 

From Consulting to Co-founding Tekkare: A Journey of Rigor and Technical Passion

 

A business school graduate, Robin Sarfati began his career in consulting, working at the crossroads of strategy and technological innovation. The turning point came during an analysis mission for the acquisition of a pharmaceutical lab by an investment fund.

 

“I was in charge of processing public data for the analysis — all in Excel at the time,” he recalls. “While exploring those sources, I discovered a fragmented world with immense potential.” This first exposure to open data sparked his passion for data processing and gradually introduced him to technical work.

 

“You really learn by doing,” Robin explains. “I was the only one able to do it, so I learned as I went. When the company grew, I delegated what I knew well and learned what I was missing.” A hands-on approach now fueling a clear mission.

 

A Clear Mission: Making Health Knowledge Accessible

 

Tekkare aims to deliver health knowledge to users in the simplest way possible. To do so, the company relies on its platform, OIP (Open Innovation Platform), now one of the largest structured public health databases.

 

“Our work involves a lot of cross-disciplinary expertise and tech,” Robin explains. “We bring together data in vastly different formats, with definitions that vary from one country to another.”

 

The stakes are high: instead of browsing various sites to get information about a drug, its consumption, or related research, everything is centralized. “For each topic, you might find three, five, six, seven different sources. It’s a huge job,” he points out.

 

Curation and ontology work enables interoperability with clients' information systems, whether they are public institutions or pharmaceutical companies. “We now serve over 30 labs, which creates real economies of scale because we only need to do this work once.”

 

CTO: A Role that Evolves with the Company’s Growth

 

For Robin, the CTO role is tightly linked to the company’s maturity stage. “In the early days, you have to code more, set the example, make it work, and make it scalable,” he explains. The goal is to build an architecture that can grow without needing a complete overhaul in two years.

 

His business school background gives him a particular edge: “System adaptability. Some CTOs are too focused on the tech, but the ability to respond quickly to client needs often matters more than building the perfect system right away.”

 

As the company grows, the role shifts toward more management and diplomacy: “You're the bridge between business needs and technical execution. You need to act as a diplomat between tech and business so both sides understand each other.”

 

Today, his mission includes strategic partnerships: “The fact that I have technical expertise and can design architecture reassures clients and helps form partnerships,” he notes.

 

Client Responsiveness: A Key Differentiator

 

This adaptable approach translates into a responsiveness that clients notice. Robin shares how he’s sometimes switched technologies mid-project to better meet client needs: “I started with one technology, but realized it wouldn’t allow us to react fast enough. So I pivoted to another, more suitable one to be ready by the next day.”

 

This responsiveness is driven by a philosophy: “We put client satisfaction at the core of our system. We understand them, strive to understand them, and design more flexible systems.”

 

Quality and Security: The Need for Certifications

 

To ensure the quality, reliability, and security of its data, Tekkare emphasizes rigorous sources and protocols, supported by certifications. “We pursued ISO 27001 and HDS certification,” Robin explains.

 

While demanding, these certifications are a strategic investment: “Now when we're asked for cybersecurity documentation, we can respond much faster. We’ve implemented all the necessary documentation, protocols, and controls.”

 

AI and Agents: A Structured, Pragmatic Approach

 

Artificial intelligence at Tekkare follows a pragmatic approach. “We don’t do AI for the sake of it,” Robin insists. “For a long time, it wasn’t relevant given the state of our data.” The available data was too fragmented to effectively train models.

 

The arrival of LLMs changed the game by enabling the exploration of the structured data ecosystem the team built. However, even though AI is now part of Tekkare’s roadmap, it’s never adopted just because it’s trendy. Robin emphasizes: “AI only makes sense if the data is structured and the need is clear.”

 

This pragmatic view reflects Tekkare’s methodical stance: no tech tools are implemented unless they deliver real added value for the client.

 

For a long time, AI wasn’t a priority at Tekkare because the conditions weren’t right. Client needs focused more on data structuring, analysis, and integration into internal tools than on automation or abstract AI usage.

 

Today, with better-structured data, AI is gradually finding its place: “We’re getting more requests, particularly from the ERDERA consortium. But we don’t hesitate to challenge the AI need. Sometimes, we tell labs they can’t do AI yet because their data isn’t structured.”

 

Still, Robin stresses that every AI project must begin with a clear framework. You must first define an objective, model the necessary data, and determine what the AI should and shouldn’t do.

 

Regarding generative AI in healthcare, Robin remains cautious: “There are still many unknowns. We don’t know the costs yet, or whether we’ll make it accessible enough.”

 

Precision remains a central concern: “For highly specific operations in healthcare, are today’s LLMs the right technology?” This question shapes Tekkare’s development: “We’re working to integrate verification mechanisms into our structured data to strengthen reliability.”

 

The example of rare diseases highlights this requirement: if an AI needs to retrieve information about a condition like Huntington’s disease, it must be connected to a rigorous, structured database. A general-purpose AI tool might not distinguish between classic Huntington’s and the juvenile form — a critical difference with major implications for genomics and treatment pathways. Using Tekkare’s data ecosystem, AI gains immediate access to detailed scientific information, enabling it to navigate the medical knowledge base accurately and deliver more reliable, relevant analyses.

 

Defining the AI Agent at Tekkare

 

For Robin, an AI agent is “more a way of doing things than a fixed concept. It’s about framing AI for a specific use case: you give it a goal, rules, and tools to use.”

 

Current projects focus on accelerating the understanding of public data and, in some cases, applying these methods to clients’ internal data.

 

Advice for Students

 

For those who want to work at the intersection of tech and healthcare, Robin recommends “a lot of perseverance and patience. Healthcare is a hard sector to master. So is tech.”

 

His practical advice: “Try to find projects that excite you — personal projects. They always push people to think, and if it excites you, you’ve probably found a good use case.” This approach helps develop market understanding while offering a concrete framework for learning.

 

“There’s nothing better than getting your hands dirty to learn and prove you know what you’re doing,” he concludes. That, he says, is how you build real expertise.