The Semenov Lab is a rapidly expanding program at the intersection of artificial intelligence, oncology, and dermatology, focused on predicting cancer outcomes and treatment toxicity through large-scale, multimodal data integration.
Based at Massachusetts General Hospital, Harvard Medical School, and Dana-Farber Cancer Institute, the lab integrates clinical data, digital pathology, and spatial profiling across one of the largest melanoma and immunotherapy cohorts globally.
Our goal is to develop scalable, data-driven approaches that directly inform clinical decision-making and improve patient outcomes.
We are building a program spanning multimodal AI (clinical + imaging + spatial data), melanoma prognostication and risk stratification, immune-related adverse events (irAEs) as biomarkers and therapeutic targets, and tumor microenvironment modeling through spatial and computational approaches. This work is supported by large, multi-institutional datasets and close integration with clinical and translational research programs.
Tens of thousands of patients across linked clinical datasets. Large, deeply phenotyped immunotherapy and melanoma cohorts with integrated histopathology and spatial imaging platforms.
Publications in Nature Medicine, Lancet Oncology, JAAD, and JAMA Dermatology. AI models that outperform existing clinical frameworks, aligned with clinical decision-making and therapeutic strategies.
Embedded within MGH and Dana-Farber melanoma and immunotherapy programs. Active collaborations across Harvard, Broad Institute, and national consortia, with integration into clinical trials.
Actively expanding AI infrastructure, national and international collaborations, and translational research. Joining now provides the opportunity to shape the next phase of program growth.
We recruit across three tracks — and welcome hybrid profiles at any intersection.
PhD-level researchers in machine learning, computer vision, or multimodal AI. Experience with medical imaging, survival modeling, or large-scale clinical data is a plus. You will lead model development for cancer prediction and digital pathology.
MD, MD-PhD, or clinician-researcher with interest in oncology, dermatology, or immunology. You will bridge clinical insight and computational methods — leading studies from hypothesis to publication to clinical translation.
Quantitative scientists with backgrounds in biostatistics, bioinformatics, epidemiology, or data engineering. You will build and validate models on one of the richest clinical + genomic + imaging datasets in academic medicine.
Opportunity to lead high-impact, first-author research. Our track record: 44+ mentored first-author publications in Nature Medicine, Lancet Oncology, JAMA Dermatology, JAAD, and npj Digital Medicine. We prioritize trainee-led work.
Access to large-scale, multimodal clinical datasets not widely available elsewhere — one of the largest linked melanoma + irAE registries globally, with 24,000+ immunotherapy patients, multi-institutional digital pathology cohorts, spatial imaging, and genomic data.
Mentorship in both computational and clinical research pathways. Alumni have gone on to faculty positions, industry roles, and NIH-funded independent careers. We tailor mentorship to academic, translational, and industry-facing goals.
We emphasize collaboration across disciplines, intellectual rigor, translational impact, and long-term career development. Exposure to major national and international meetings and conferences.
We have openings for postdoctoral fellows, research scientists, and research fellows. All positions offer protected research time, mentorship, and a clear path to first-author publication.
We are seeking a postdoctoral fellow with a strong background in machine learning, computer vision, or computational biology to advance our AI in dermatology and digital pathology pipeline for melanoma prognosis. This position involves working with foundation models (Virchow2), whole-slide image analysis, and multimodal data integration across clinical and genomic datasets in cutaneous oncology.
We welcome medical students and recent graduates interested in gaining hands-on research experience in clinical informatics in melanoma, AI in skin disease, and oncodermatology. Fellows contribute to ongoing projects in AI in melanoma prognosis, irAE surveillance, immune-mediated skin disease, and clinical trials in oncodermatology. Prior research experience and interest in academic medicine are preferred.
We collaborate closely with students in the Harvard Department of Biomedical Informatics (HMS DBMI). If you are admitted to or enrolled in the HMS DBMI program and are interested in dermatology AI, melanoma ML, or clinical NLP, please reach out to discuss potential research projects and thesis opportunities.
I am always excited to meet people who are passionate about applying rigorous science — AI in dermatology, clinical informatics in melanoma, or cutaneous oncology — to improve outcomes for patients with skin cancer, immune-mediated skin disease, and inflammatory skin conditions from cancer therapy.
To apply, send an email to ysemenov<at>mgh.harvard.edu with the subject line "[Position] Application — [Your Name]" and include:
