Kinan Dak Albab Joins CDS as Assistant Professor, Advancing Research and Education in Data Privacy
In fall of 2025, Boston University’s Faculty of Computing & Data Sciences (CDS) welcomed Kinan Dak Albab as an Assistant Professor. With expertise at the intersection of data systems, programming languages, cryptography, and privacy law, Kinan’s work focuses on developing practical, interdisciplinary solutions to safeguard digital privacy in real-world systems.
“Data privacy is one of the defining challenges of our time, and Kinan brings both the technical mastery and the cross-disciplinary perspective needed to address it,” said Azer Bestavros, Associate Provost for Computing & Data Sciences. “His research spans engineering, law, and societal implications—exactly the kind of holistic thinking that distinguishes CDS. I’m excited for the impact he’ll have on our students and the field.”
For Dak Albab, CDS offers the ideal environment to pursue both pillars of his career: research and teaching. “CDS is a hub that brings disciplines together,” he said. “To make a real impact on digital privacy, you have to think about it holistically—across technology, user expectations, industry incentives, and policy. And teaching is just as important—equipping students from all backgrounds with the tools and mindset to keep privacy at the forefront of their work.”
“CDS is a hub that brings disciplines together… To make a real impact on digital privacy, you have to think about it holistically.”
At CDS, Dak Albab is teaching an innovative, project-based course on data systems for privacy, alongside his ongoing research on privacy infrastructure, regulation, and system design. He was selected to receive one of the 2025 Shipley Center Academic Innovation Awards in support of the development of “PaperBuddy” – a system that leverages AI to assist students with reading assignments, focusing on the intentional and structured integration and use of LLMs into flipped classroom settings to improve student satisfaction and performance. In this Q&A, he shares what excites him most about joining CDS, his holistic approach to privacy, how he empowers students to lead their projects, and his advice for those interested in interdisciplinary collaboration.
Q&A
What excites you most about joining CDS and this new role?
I decided to pursue a career in academia for two reasons: research and teaching/mentorship. I am excited to join CDS for both reasons.
In terms of research, I would like to have a real impact on individuals' digital privacy in the real world. To do so, I need to think about privacy problems holistically and through many lenses. It is not enough to simply come up with technical systems or mathematical formulas that exhibit or define certain privacy properties. We also have to ask ourselves many other questions. Are these the "right" privacy properties (and what does right mean)? Do they align with the expectations of end users? Do they actually achieve their guarantees in practice? How can we align them with existing incentives in the industry? These are broad questions that span many fields of inquiry. They connect classical technical topics from data science and computer science with law, society, and psychology. CDS is a hub that brings these disciplines together.
Equally important is teaching. One of the best ways to make a real impact on privacy is to teach cohorts of students to think about privacy questions and equip them with the right tools and resources for that thinking. This includes people who go on to be software or data engineers who build pipelines for analyzing data, and those who go on to research, data science, or the public sphere (e.g., journalism, policy, etc.). I was impressed by CDS's focus on data ethics and the impact that hit ad even beyond CDS majors. I hope to replicate some of that success with data privacy.
As a researcher joining CDS, how do you see data science enhancing the impact of your work?
Data science is a growing field where there is a significant tension with privacy. One can view privacy as a concern that restricts what kind of analytics is allowed to be executed or revealed about data (or how those analytics are carried out).
In a way, this implies that no "privacy solutions" would get broadly adopted unless data scientists "like them" (a necessary but insufficient condition). Our notions of privacy and our tools and systems for realizing these notions must be strong enough to provide reasonable protections to end users, but permissive enough to let data scientists carry out their core, reasonable functions without too much burden.
Being at CDS and being close to data science and data scientists (at all levels) gives me a unique perspective into how to best achieve this balance, as well as how to empirically evaluate where solutions land on this balance.
Can you highlight one project that best represents your work?
I am investigating how to build new infrastructure systems that assist data engineers and software developers in meeting their desired privacy policies. These include databases, web frameworks, and microservice systems. A key component of this work relies on existing privacy regulations (and the threat of fines) to incentivize rational behavior from these engineers. This rationality assumption allows building lighter-weight systems with good performance and limited burden to engineers. My two most recent papers, K9db and Sesame, fall in this direction, as well as several ongoing projects.
This direction combines several areas and techniques that have been close to my heart since I first became interested in computers: data systems, programming languages, and cryptography. It also has an interesting interplay with privacy laws, including the potential to inform the interpretation of these laws (and future laws). For example, by demonstrating which "privacy definitions" are and are not possible for engineers to comply with, and the degree of burden or error-proneness that different definitions impose.
You’ll be teaching an innovative course with student-led projects—what’s the vision behind that?
I think the best way for students to learn is to do things themselves. I have always preferred project-based courses to exam-based ones for that reason --- even when I was a student.
The course I am teaching looks at privacy (and specifically data systems for privacy) via the same holistic lens I outlined above. This lens covers many radically different challenges that often require different (but often complementary) solutions.
By allowing students to propose their own project ideas (with plenty of guidance, of course!), I give students the freedom to choose which aspects of this holistic picture to focus on and what to prioritize, in a way that they feel is best suited for their own interests, goals, and backgrounds.
This is not merely a nice sound bite; I have some evidence to prove it. I taught three iterations of this course at Brown before joining CDS, all sharing this philosophy. In each of these iterations, the students and I ended up with a wide cohort of ambitious projects, ranging from classic data systems design and implementation to applied cryptography, programming languages, statistical data analyses, user and legal studies, and beyond. In fact, one of my favorite parts of the course is the project presentations at the end of the course, where we get to see a microcosm of the diverse privacy challenges and solutions in the real world!
Perhaps the most important thing is that my past students must have thoroughly enjoyed this approach and found it rewarding. After all, many ended up doing research and publishing papers about related topics with me or with others!
How will students benefit from working on a project in your class?
In general, students will get hands-on experience with academic research around problems in data systems and in privacy. They will hone their design and implementation skills of complex data systems, and develop an appreciation and an intuition for why data systems are designed the way they are, which often allows them to utilize these systems more effectively. Students will also learn to design experiments and conduct systematic evaluation of their systems, especially for run-time performance on representative loads. Students also acquire a decent amount of subject matter expertise in the area of their project of choice (e.g., privacy regulations, information flow control, etc.).
How do your research and teaching audiences intersect?
There is a close relationship between the two. My research is focused on building privacy systems and abstractions that are easy to use by data scientists and engineers. This audience overlaps with the students that I will be teaching.
On the whole, one way to view this course is that it is an advert for data + privacy research. The goal is to survey the research space for students and allow them to dive deeper into a corner of it that they chose. I hope that this would keep privacy concerns at the top of their minds later in their careers and equip them with the resources they need to identify these problems, and existing research and tools they could employ to mitigate them effectively.
You’ve engaged in interdisciplinary collaboration—what advice do you have for prospective students?
Exploration is a key part of learning and the student experience. This necessarily includes some exploration that leads nowhere, either because it was too ambitious, too unfocused, or led in a direction that became uninteresting to the student. That is fine and expected. I do not believe that any time should be considered wasted when it leads to learning valuable lessons (even "negative results").
I think a key part in making an interdisciplinary collaboration effective is mentorship and guidance, ideally by several people who complement each other and together have expertise that spans the different disciplines. Another key part is balancing depth and breadth. Acquiring breadth by dipping one's toes in different areas is very helpful for interdisciplinary collaboration. It allows students to see the big picture and the concrete and less-concrete connections. However, it is important that students eventually focus on a particular problem that they find interesting and dive into it.
Beyond the Bio: What do you enjoy outside the classroom?
I am really into cooking and cocktail making, and I think I have gotten pretty good at it. The readers can judge for themselves: I like to cook at least one group meal for my students every semester! We cannot do research on an empty stomach.
There are other personal quirks that I could -- but should not -- share with you. One of the assignments in my course motivates a statistical approach to privacy called "differential privacy". In this assignment, I ask students to find out personal information about me by performing various statistical analyses and attacks and correlating them with a trail of information I left online, so I cannot give these answers out now.
By Maureen McCarthy