{"$schema":"https://raw.githubusercontent.com/jsonresume/resume-schema/v1.0.0/schema.json","basics":{"name":"Mohammad Haft-Javaherian","label":"Managing Member, Haft Advisors","image":"https://saywise-production-profilepicturestoragebucket-kbfooccc.s3.amazonaws.com/profile-pictures/e51ab9e6-a96c-4754-9600-b9f9de2d4b7b/linkedin-1791995914989.jpg","summary":"AI leader with a PhD in biomedical engineering who has built and led machine learning teams developing FDA-approved medical AI for clinical trials and GI endoscopy. Now runs a boutique consulting firm applying computer vision and LLMs in medical device and healthcare.","location":{"city":"US"},"profiles":[{"network":"LinkedIn","username":"mohammadhaft","url":"https://linkedin.com/in/mohammadhaft"},{"network":"Google Scholar","username":"shyv6tAAAAAJ","url":"https://scholar.google.com/citations?user=shyv6tAAAAAJ"}]},"meta":{"canonical":"https://saywise.com/member6386","version":"v1.0.0","lastModified":"2026-10-14T16:38:36.959Z"},"x_saywise":{"handle":"member6386","pronouns":null,"availability":null,"oneLiner":"AI leader with a PhD in biomedical engineering who has built and led machine learning teams developing FDA-approved medical AI for clinical trials and GI endoscopy. Now runs a boutique consulting firm applying computer vision and LLMs in medical device and healthcare.","profileUrl":"https://saywise.com/member6386","markdownUrl":"https://saywise.com/member6386/profile.md","pdfUrl":"https://saywise.com/member6386/resume.pdf","archetype":{"code":"HS-MX","name":"Launchpad Artisan","url":"https://saywise.com/sca/215427bf167a","rarity":"One of the first Launchpad Artisans on Saywise"}},"work":[{"name":"Haft Advisors","position":"Managing Member","location":"Irvine, California, United States","url":"https://haftadvisors.com","startDate":"2026-04-01","summary":"Boutique consulting firm applying computer vision and LLMs in medical device and healthcare. Provides fractional head of tech services to early-stage startups and corporations, focusing on integrating AI into workflows with an emphasis on business outcomes. Develops strategic plans to improve operational efficiency and drive tech adoption."},{"name":"Iterative Health (Formerly Iterative Scopes)","position":"Head of Machine Learning | Director of Machine Learning","location":"Cambridge, Massachusetts, United States","startDate":"2021-04-01","endDate":"2026-03-01","summary":"Oversaw the machine learning department with up to four manager reports and 28 professionals, promoted from Manager of Machine Learning to Head of ML reporting to the CEO. Led multiple generations of FDA-approved AI colorectal cancer screening devices, developed a GI foundation model, and owned pharma and Eli Lilly contracts. Established peer review and SDLC practices, cut deployment time from two months to two weeks, and built hiring and career ladder programs."},{"name":"Iterative Health","position":"Manager of Machine Learning R&D | Senior Machine Learning Scientist","location":"Cambridge, Massachusetts, United States","url":"https://iterative.health","startDate":"2021-04-01","endDate":"2022-05-01","summary":"Here, I oversaw a team of researchers and engineers developing ML technologies and evolving foundational research to optimize clinical trials. ✦ 2X development speed. Led design and implementation of end-to-end production-level pipeline for training, inference, and automatic orchestration, including model registry, experiment tracking, CI / CD, and ML operations (cloud, edge). ✦ Set up ML governance, model interpretability / explainability (XAI), ethical AI / bias mitigation, and scalable data / cloud strategy."},{"name":"Harvard Medical School and Massachusetts General Hospital","position":"Researcher Fellow","location":"Cambridge, Massachusetts, United States","startDate":"2019-05-01","endDate":"2021-04-01","summary":"Developed a novel machine learning method for cardiovascular OCT images using CNNs, a Wasserstein-based critic model, and a novel topological loss. Reduced error 77% (30 to 7 microns) and developed a CNN process for intracoronary PS-OCT 3D images."},{"name":"Cleerly","position":"Intern Data Scientist","location":"New York City","url":"https://cleerlyhealth.com","startDate":"2018-05-01","endDate":"2018-08-01","summary":"I contributed to the design and development of an AI-based (deep learning, computer vision) calcium score and coronary artery ischemia prediction models in cardiac CT images. ✦ 7X increase in detected adverse cardiovascular events during 8-year follow-up."},{"name":"New York-Presbyterian Hospital","position":"Immersion Internship","location":"New York","startDate":"2016-06-01","endDate":"2016-08-01","summary":"I worked in MRI imaging, diagnosis, and multiple sclerosis outpatient clinic observerships."},{"name":"Cornell University","position":"Graduate Research And Teaching Assistant","location":"Ithaca, New York Area","startDate":"2015-08-01","endDate":"2019-05-01","summary":"I studied brain vasculature using computational fluid dynamics and mechanics. ✦ Decreased analysis time from 20 hours to just a few minutes by developing ML models (CNN and RNN); highly cited and adopted by several labs worldwide. ✦ Attracted 57K+ users with automated crowdsourcing pipelines developed for Alzheimer's research."},{"name":"Cornell University","position":"Graduate Teaching Assistant","location":"Ithaca, New York, United States","startDate":"2015-01-01","endDate":"2015-05-01","summary":"I taught programming labs and held office hours to support students in the School of Computer Science."},{"name":"Cornell University & Other Institutes","position":"Graduate Research & Teaching Assistant","location":"Ithaca, NY","startDate":"2006-01-01","endDate":"2019-05-01","summary":"Graduate research and teaching assistant across Cornell University and other institutes during doctoral work in biomedical engineering."}],"education":[{"institution":"Massachusetts Institute of Technology","studyType":"Postdoctoral Fellowship in Computer Science","startDate":"2019-05-01","endDate":"2021-06-01"},{"institution":"Cornell University","url":"https://cornell.edu","studyType":"Doctor of Philosophy (PhD) in Biomedical Engineering","startDate":"2013-09-01","endDate":"2019-06-01"},{"institution":"University of Tehran","studyType":"Bachelor of Science (BS) in Civil Engineering","startDate":"2005-09-01","endDate":"2009-06-01"}],"skills":[{"name":"Machine Learning"},{"name":"Deep Learning"},{"name":"Computer Vision"},{"name":"Convolutional Neural Networks (CNNs)"},{"name":"Recurrent Neural Networks (RNNs)"},{"name":"Generative Adversarial Networks (GANs)"},{"name":"Large Language Models (LLMs)"},{"name":"Python"},{"name":"Matlab"},{"name":"C++"},{"name":"Fortran"},{"name":"PyTorch"},{"name":"TensorFlow"},{"name":"Scikit"},{"name":"OpenCV"},{"name":"SQL"},{"name":"NoSQL"},{"name":"Jira"},{"name":"GitHub"},{"name":"Jenkins"},{"name":"Kubernetes"},{"name":"Agile Development"},{"name":"Technology Management"},{"name":"Business Strategy"},{"name":"Managing the Learning Function"},{"name":"Multi-agent Systems"},{"name":"Regulatory Submissions"},{"name":"Leadership"},{"name":"C-Level Communications"},{"name":"Artificial Intelligence (AI)"},{"name":"Strategic Vision"},{"name":"Cross-functional Collaborations"},{"name":"Influencing Without Authority"},{"name":"Translating Strategy"},{"name":"Technological Innovation"},{"name":"Adaptive Leadership"},{"name":"Mission Oriented"},{"name":"Data-driven Decision Making"},{"name":"Team Building"},{"name":"Empowerment"},{"name":"Delegation"},{"name":"Operations Process Improvement"},{"name":"Conflict Management"},{"name":"Change Management"},{"name":"Performance Improvement"},{"name":"Empathy"},{"name":"Social Awareness"},{"name":"Emotional Intelligence"},{"name":"Active Listening"},{"name":"Collaborative Problem Solving"}],"certificates":[{"name":"Advanced Leaders Program","date":"2023-01-01","issuer":"Eager Labs"},{"name":"Management Training","date":"2023-01-01","issuer":"LifeLabs Learning"},{"name":"Connect for Developing Leaders","date":"2022-08-01","issuer":"Eager Labs"},{"name":"Colman Leadership Program","date":"2017-01-01","issuer":"Cornell University"},{"name":"Lean Six Sigma Black Belt (ICBB)","date":"2016-01-01","issuer":"The State University of New York"},{"name":"Science Communication","date":"2016-01-01","issuer":"Cornell University"},{"name":"Leadership Assessment for Managers","date":"2015-01-01","issuer":"Cornell University"},{"name":"Leadership Assessment for Managers","date":"2015-01-01","issuer":"Cornell University"}],"publications":[{"name":"Summary","summary":"Google Scholar page’s link: 1,297 citations, h-index of 14, & 45 publications Highly cited publications in Biomedical AI/Computer Vision (1,297 citations and h-index 14 with 45 publications)"},{"name":"A Topological Encoding Convolutional Neural Network for Segmentation of 3D Multiphoton Images of Brain Vasculature Using Persistent Homology","summary":"The clinical evidence suggests that cognitive disorders are associated with vasculature dysfunction and decreased blood flow in the brain. Hence, a functional understanding of the linkage between brain functionality and the vascular network is essential. However, methods to systematically and quantitatively describe and compare structures as complex as brain blood vessels are lacking. 3D imaging modalities such as multiphoton microscopy enables researchers to capture the network of brain vasculature with high spatial resolutions. Nonetheless, image processing and inference are some of the bottlenecks for biomedical research involving imaging, and any advancement in this area impacts many research groups. Here, we propose a topological encoding convolutional neural network based on persistent homology to segment 3D multiphoton images of brain vasculature. We demonstrate that our model outperforms state-of-the-art models in terms of the Dice coefficient and it is comparable in terms of other metrics such as sensitivity. Additionally, the topological characteristics of our model's segmentation results mimic manual ground truth. Our code and model are open source at https://github.com/mhaft/DeepVess."},{"name":"Brain capillary networks across species: a few simple organizational requirements are sufficient to reproduce both structure and function"},{"name":"Deep convolutional neural networks for segmenting 3D in vivo multiphoton images of vasculature in Alzheimer disease mouse models","summary":"The health and function of tissue rely on its vasculature network to provide reliable blood perfusion. Volumetric imaging approaches, such as multiphoton microscopy, are able to generate detailed 3D images of blood vessels that could contribute to our understanding of the role of vascular structure in normal physiology and in disease mechanisms. The segmentation of vessels, a core image analysis problem, is a bottleneck that has prevented the systematic comparison of 3D vascular architecture across experimental populations. We explored the use of convolutional neural networks to segment 3D vessels within volumetric in vivo images acquired by multiphoton microscopy. We evaluated different network architectures and machine learning techniques in the context of this segmentation problem. We show that our optimized convolutional neural network architecture with a customized loss function, which we call DeepVess, yielded a segmentation accuracy that was better than state-of-the-art methods, while also being orders of magnitude faster than the manual annotation. To explore the effects of aging and Alzheimer’s disease on capillaries, we applied DeepVess to 3D images of cortical blood vessels in young and old mouse models of Alzheimer’s disease and wild type littermates. We found little difference in the distribution of capillary diameter or tortuosity between these groups, but did note a decrease in the number of longer capillary segments (>75μm) in aged animals as compared to young, in both wild type and Alzheimer’s disease mouse models."},{"name":"Neutrophil adhesion in brain capillaries reduces cortical blood flow and impairs memory function in Alzheimer’s disease mouse models"},{"name":"A Numerical - Experimental Approach to Characterize Fracture Properties of Asphalt Mixtures at Low In - Service Temperatures"},{"name":"Two-Dimensional Virtual Microstructure Generation of Particle-Reinforced Composites","summary":"Abstract This paper describes the development and validation of a virtual microstructure generator that can be used to fabricate a virtual, two-dimensional (2D) microstructure of particle-reinforced composites in general, and of asphalt concrete and/or cement concrete mixtures in particular. In the generator, some core geometric characteristics such as aggregate gradation, aggregate volume fraction, and angularity were used to transform data from a three-dimensional (3D) mixture into its 2D microstructure. Resulting virtual ..."},{"name":"Microstructure modeling of rate-dependent fracture behavior in bituminous paving mixtures","summary":"This paper presents a computational microstructure model to predict the fracture-related behavior of heterogeneous and viscoelastic bituminous paving mixtures. In addition to the consideration of the complex geometric characteristics and viscoelastic behavior of the mixture, this study implements a rate-dependent cohesive zone model into the mainframe of ABAQUS in the form of a customized user element (UEL) subroutine. This enables model fracture as a gradual and rate-dependent phenomenon in which the initiation and ..."},{"name":"Modeling Rate-Dependent Fracture Behavior of Asphalt Mixtures","summary":"Abstract: This paper presents a computational microstructure model to predict the fracture- related behavior of heterogeneous and inelastic asphalt mixtures. In addition to the consideration of the complex geometric characteristics and inelastic behavior of the ..."},{"name":"Virtual Fabrication Of Composite Mixtures: In Civil Engineering","summary":"This book describes the development and application of a method of virtual generation incorporated with post-processing image analysis that can be used to fabricate a virtual, two-dimensional microstructure of composite mixtures. In the generator, geometrical characteristics such as aggregate gradation, aggregate area fraction, angularity, orientation, and elongation are used to transform data from a three-dimensional image into a two-dimensional image.The generator can provide significant savings in time and cost, and it can also be used to improve current methods for the mechanical characterization of composite mixtures."},{"name":"A Cohesive Zone Fracture Model to Predict the Rate-Dependent Fracture Behavior of Heterogeneous and Viscoelastic Asphalt Mixtures"},{"name":"Modeling to Simulate Fracture Damage in Heterogeneous and Inelastic Asphalt Mixtures Based on Rate-Dependent Cohesive Zone Fracture"},{"name":"Virtual Microstructure Generation of Asphaltic Mixtures","summary":"Abstract This thesis describes the development and application of a virtual microstructure generator incorporated with post-processing image analysis methods that can be used to fabricate a virtual, two-dimensional microstructure of asphaltic mixtures. In the generator, geometrical characteristics such as aggregate gradation, aggregate area fraction, angularity, orientation, and elongation were used to transform data from a three-dimensional (3D) mixture into its two-dimensional (2D) microstructure. The 2D virtual microstructures were ..."}]}