Piotr Orlowski Email & Phone Number

Cambridge, Gb

Developing an ultrasound-based device for facilitated and more accurate measurement of heart performance signals- Prioritised product portfolio based on a balance of feasibility, product market fit and envisaged ROI. Secured technology strategy buy-in from technical, financial and medical co-founders - Developed the product technology concept and proof of concept (PoC) activity roadmap in partnership with clinical co-founders. Short-term focused on cashflow generating product yet with significant barrier to entry; long-term anticipating evolution of trends in AI / robotics / augmented reality- Directed contract team of 3 engineers + own development (software and hardware) to build the prototype of device electronics for augmented probe orientation characterisation- Developed prototype computer vision model for real-time pose-based patient positioning - Developing feedback control between ultrasound images and probe position to guide probe movement - Arranging test data acquisition on initial prototype with an academic partner- Authored technical documentation and delivering development vision to prospective investment partners - Setting-up foundational product development infrastructure (Arduino / C++ hardware control; AI (PyTorch, Tensorflow, OpenCV) + MLOps; VisionOS for Augmented Reality); Python package templates, GitHub Actions

Data Engineering, DataOps and Governance consulting for Oncology Data Science at AstraZeneca- Delivered 4 data engineering prototypes in support of the scale-up of the group’s operations: 1) Data Quality Control python package for detecting and flagging entry type, format and control vocabulary errors based on concepts in data rather than defining checks on a column-by-column basis for each dataset as part of DataOps; Projected ~80% reduction of developer time for this type of ETL. Deployed as part of an E2E automated ETL pipeline from AWS S3 to 3rd party analysis platform,2) A Python template for standardising development of data engineering packages (cookiecutter) adopted as standard by team and shared with leadership for wider adoption, cut ~2 weeks setup time per project standardisation of packages to simplify CI / CD pipelines;3) Python package for formatting vendor clinical data input Excel files to a standard based on config describing the sequence of operations,4) data pipeline that mapped vendor shipment manifests for ctDNA samples (40 000 samples across 8 data providers and 3 internal storage systems) to an internal standard in support of Oncology R&D- 2nd place out of 23 teams for internal innovation project on FAIR data at source project leveraging a novel knowledge management standard linkML to apply data modelling rules to ontology-based concepts and enable the rules to be defined by natural language and translated into code through generative AI;- Coached junior director in the setup of a Data Operations function at the interface of Data Engineering, Data Governance, Product and Business stakeholdersExploring new business opportunities:- Creation of mathematical models of injury for expert advice in legal cases for medical malpractice - Optimisation of grassroot innovation events in pharma to maximise impact (with AstraZeneca and Cambridge)

Brentford, Middlesex, Gb

- Setup a team of 2 product owners, 2 data engineers, a programme manager and a metadata lead to prototype and own data products in support of oncology biomarker and companion diagnostics (CDx) research programmes- Set a strategy by prioritising 15 products focused on speed, quality and interoperability data improvements for precision medicine applications. Aligned Research, AI, Engineering and Data Management stakeholders to support the portfolio- Data products delivered: 1) replaced hard drive data shipment from a vendor with a remote transfer service, 2) automated with JIRA / Confluence the tracking of clinical data ingestions and transformations into GSK, 3) developed a prototype for automatically generating data transfer agreements, 4) developed a prototype data quality control (QC) service for 5 bioinformatics data types, 5) developed a data transformation service for transfer of data into 3rd party software, 6) developed data governance guidance for use of a large (20-300TB) academic dataset by AI and R&D stakeholders 7) developed a flow cytometry data and metadata model in view of modelling data for deep phenotyping applications

Brentford, Middlesex, Gb

Clinical Data- Made a successful case for investment in establishing a data science group (11 people) for optimisation of clinical trial operations. Structured the team in Data Engineering, Data Science, Data Strategy and Visualisation units- Led the transfer of work from a management consulting team to the new group . The team supported 6 clinical trials and developed requested tools and insights on time. A model and dashboard predicting the change in seasonality of a respiratory disease during COVID lockdown and thus availability of patients for a trial was the highlight result that allowed the optimal start time for a trial to be selected. Stack: Python, R, Power BI, SQL / Hadoop / Impala- Aligned data engineering and data governance teams to address in 2 weeks a systemic data security risk that stopped all analysis of our team and clients. Managed customer expectations during this uncertain time. This work led to a change of standard of transfer of data by our clinical operations customers - Led the transition of the team from an independent squad to a squad member of a tribe aiming to develop a data science and visualisation platform for planning clinical trials. Set the Data Science agenda for the teamImaging data- developed a community of 40 imaging experts to define imaging needs at GSK and prioritise the group’s work- Summarised and extended gathered insights into an imaging data strategy and prioritised the development of a digital pathology pipeline as first step of the execution of the proposal based on the generic nature of many of its components for other pipelines and strategic importance of pathology for GSK’s investment in Oncology- Evaluated external contractors for the development of pre-processing components of the imaging pipeline. Evaluation led to the resourcing of one contractor for improving data acquisitions and me developing a software package for detecting 3 artifacts in the data. Currently used in the pathology department

Brentford, Middlesex, Gb

- Developed a prototype AI tool for identifying ovarian cancer drug responders from Digital Pathology and Bioinformatics data by leading a team of 3 data scientists. Work led to continued investment in 6 AI engineers to develop the product. Stack: Python, Pytorch, OpenCV, bash, Slurm- Prototyped a dashboard for AI-based optimisation of a biopharmaceutical process (reduction of critical path by 3 months) through leading a team of 4 data scientists, an external database contractor and a biologist. Tool transferred to production. Stack: Python, Plotly, bash, APIs- Setup coding standards, development processes and tools to deliver quality code fast. As a result, development time of a new model by members of my team was reduced from two weeks initially to two days when leveraging code modularity

Oxford, Oxfordshire, Gb

- Built the Data Analysis team at ONI. To this end, designed and implemented a platform for pipelining microscopy image analysis in 5 different areas (segmentation, classification, augmentation, clustering, motion)Stack: Python, JS / React, Flask- Recruited to parallel stream activity in the five areas including a top 5% Graduate Engineer from Oxford, an experienced Data Analyst from the competition, an experienced Data Scientist and a JS Developer- Through sponsorship of a scientific conference organised by the client interested in the data analysis product advertised our tools to the broad community (200 people) of potential users for refinement and market size assessment (£2 million / year)- Negotiated pricing of the data analysis tools for a first license contract with a client. Achieved £100k, £50k above target- Led the team through a transition from a flat to a vertical and then to a matrix management structure as the company grew from ~10 employees when I joined to ~70 when I left- Presenting the R&D vision/results to prospective investors as well as to the broad company team. Results of the team contributed to a major investment (£100 million valuation) in the company which set ONI to grow to over 100 employees

Based on specifications devised with a clinical partner I developed a prototype MATLAB software for the quantification of tissue content around coronary arteries from Computed Tomography Angiography images. Steps of analysis included: segmentation of the coronary tree, vessel network extraction, vessel segment of interest selection via a GUI, vessel straightening, computation of geometric, contrast and tissue metrics along the vessel, data export and visualisation of steps of the analysis within a GUI

- As head of the Imaging Research Cluster at the college I initiated and built 5 multidisciplinary research teams involving engineers, clinicians and biologists focused on imaging and modelling the pathophysiology of brain vascular conditions including stroke, multiple sclerosis and neuromyelitis optica. This work led to identifying new therapeutic targets estimated effective drug dosage and eliminated some competing hypotheses about the mechanisms behind disease progression. - I recruited a team of 6 students (5 MEng and 1 PhD). I managed my team for maximum group output and for the individual growth of its members. All of my students were invited to present their research at international conferences and 2 received invitations for full paper journal publications (best 20% of papers). My PhD student graduated with 2 articles published

Oxford, Oxfordshire, Gb

- Tutoring and delivering revision classes in 1st and 2nd year Mathematics and Electronics (3 h/week)- Co-directing studies and providing pastoral care for 23 undergraduate students

Oxford, Oxfordshire, Gb

- Designed physiological models of metabolism and injury of brain cells and tissue post stroke in support of creating new imaging modalities and planning stroke treatment including device design.- Developed image processing tools and worked on the spin-out of software for the planning of the treatment of arteriovenous malformations (AVMs) of the brain with clinical and industrial partners - Supervision of undergraduate and graduate research students- Co-designed and supervised of a mathematical physiology laboratory for 1st year doctoral students training in Mathematical Physiology

Oxford, Oxfordshire, Gb

- Taught Mathematics, Electronics and Energy and Environment courses (4 h/week)- Interviewed during university undergraduate admissions- Member of the College Finance Committee and involved in College fundraising campaigns

Designed parts of a magnetic rail grinding residues evacuator.

Houston, Texas, Us

Developed 2 SIM card Java Applets: 1) lost phone call number recovery 2) Car fleet management: retrieval of particular car location by SMS

Executive Mba

Doctor Of Philosophy - Phd, Engineering Science

Msc, Image And Signal Processing

Msc, Electronics And Information Technology

Ma, French