Onur

[14/09/26] Joined dsm-firmenich

I joined dsm-firmenich as an AI/ML Data Scientist to help decelerating human aging. Even though it sounds cliche, I am really excited for the opportunity as translating science into reality is my true passion. I believe with the help of AI, the opportunity is indeed here and the window is wide-open for us to grab it. Let’s see!

[16/06/26] Defended my PhD

It was quite a bit of a journey to reach this milestone. The day itself was quite special for me, as almost everyone I love and respect was there (or online) to celebrate it with me. Even though the last bits of the journey was insanely nerve-wracking experience, now I can look back at the whole thing and be content with what I have done, and how much we’ve contributed to the field as a team.

Overall, I am very satisfied with myself in terms of the science, and the potential positive impact our research might bring to patients in the future. Unfortunately, the treatment options of hematological malignancies are currently at their infancy. However, we’re seeing the first glimpses of new generation of drugs and treatment options, and how they are revolutionizing the field. I am quite certain within the next decade the field will be unrecognizable and we’ll be able to treat cancer (perhaps prevent arriving to that stage) as if it is a flu.

Personally, this whole thing was an invaluable learning opportunity. I find a chance to be in a very unique position where I had a chance to daily interact with the groups who conducted the first gene-therapy treatment in Netherlands, the first TCR-T cell therapy for NPM1-mutated AML in the world (?)… I got to sit next to perhaps top-experts on human thymus, worked with human coord-blood data (which was notoriously hard to annotate), tried to understand a human disease with cutting-edge technologies, had daily discussions with doctors & clinicians who were literally trying to save people, be criticized by the best of the best on every field I thought I knew :)

One of the funniest part of the defense (for me) was I almost got no question about computational methodologies. It was purely biology, and I happen to knew quite a bit of them to defend the work we’ve done. I have no formal education in biology, I learned quite a bit of them during meetings, heard from an expert while having a lunch together, ear-dropped a conversation while sitting in the office, but in the end I was able to answer an immunologist’s question on why IL6 promoters were chosen as a target, why we need combination therapy… to me this is really funny, because I find myself asking why would an electronics engineer know all of this? But then I remember the bar is super high, when you have people like Leonid Hurwicz, who was awarded Nobel prize in economics saying: “Whatever economics I learned I learned by listening”. So, I tell myself I could have listened better :)

The thesis can be accessed here.

[08/06/26] Presented our recent findings in Rotterdam

Shared our recent findings with the community, and talked about our new pipeline for predicting Acute Leukemia cases from transcriptomics data.

[09/01/26] Numbword was featured in PCGamer

Our daily word puzzle game Numbword was featured by PC Gamer as one of the best Wordle alternatives. We also just crossed 300K+ total plays. 🎉

[15/09/25] Our paper is published in Leukemia

We dissect NPM1-mutated AML into finer subtypes and demonstrate that these distinctions are consistent across modalities (mRNA and cell surface proteins) using single-cell technologies.

Importantly, we also identify rare hematopoietic stem cells (HSCs), both healthy and leukemic, shedding light on the earliest events in leukemogenesis with potential implications for prognosis and therapy.

Read the paper here.

Older News

[07/05/25] Launched DuoBook (#1 post on HackerNews)

We launched DuoBook, a platform for bilingual reading across the web, helping users learn languages through real content; the project reached #1 on Hacker News at launch.

[19/03/25] Our paper in Aging Cell was recognized as one of the most cited papers of the year

[22/01/25] Presented our results at DHC’2025

[25/12/24] Our latest preprint is online

[16/09/24] ceLLama is highlighted in Nature

Our single-cell annotation framework ceLLama is featured in Nature, highlighting the potential of local large language models for scalable and interpretable cell-type annotation.

[16/09/24] Our paper is published in Blood Adv.

Our work in Blood Advances investigates hematopoietic biology and regulatory mechanisms relevant to stem/progenitor cell function and therapeutic applications in blood disorders.

[18/05/24] Our paper is published in npj PO.

In npj Precision Oncology, we introduce a single-cell–guided deconvolution framework that reconstructs AML differentiation states and predicts drug response, linking cellular composition to therapy resistance.

[15/02/24] Our paper is published in Leukemia

In this study we map transcriptomic heterogeneity across AML cohorts, identifying novel subtypes with distinct biology and drug response profiles beyond genetic classification.

[07/01/24] Our paper is published in Nature Immunology

We characterize how baseline immune states shape vaccine responses, revealing distinct cellular and transcriptional predictors of immunological outcomes in aging.