Serhiy Lubensky

As the Treasurer and CFO at the Ukrainian Refugee Foundation (theURF.org), I manage the financial administration with a focus on fundraising in the United States. I ensure the smooth and cost-effective transfer of funds both domestically and internationally, leveraging technology platforms to streamline donation processing and optimize payment systems. Additionally, I oversee donor management activities, building strong relationships with supporters and ensuring that funds are securely and efficiently allocated to assist refugees, displaced families, and frontline efforts in Ukraine.

As a member of the teams integrating business intelligence assets of heritage FRB into JPMC, I have managed and contributed to the following efforts: - Re-building of ML pipelines due to changes in the technology platform and banker and client populations, product types, account usage and client engagement channels, following the acquisition of FRB. - Analytical support of HR and executive management of the affected JPMC business units, including client reassignment, product mapping, client cash flow analysis and usage stratification, etc. - GenAI prototype for automating and scaling the delivery of data-driven insights, on-demand synthesis of personal information sheets and a RAG to answer questions about bankers' book of business. - Development of a "fake" bank based on FRB statistics, from transactions to accounts, clients, bankers and branches, built in a way that's self-consistent, realistic, but untraceable to the actual bank data. - Creation and support (through semi-weekly meetings) of a Data Science Club, drawn from analysts, engineers and data scientists across multiple departments of heritage FRB taking part in JPMC transition.

First Republic Bank, reputed for delivering personalized, relationship-based service to its clients has grown at an exponential rate over the period from 2015 to 2021, resulting in some bankers' count of client relationships increasing by an order of magnitude in a short amount of time. The purpose of the newly created Eagle Intelligence department was to leverage technology to empower the bankers to scale up their high-touch service through proactive, targeted, and timely outreach to clients. As head of ML function, my job was to oversee the implementation of the intelligence pipelines in support of this goal. At its peak, the program induced over $1B in additional deposits per year, second only to the top two producer teams. It was actively used over 90% of the salesforce, consuming insights on the website, through mobile application and through targeted daily email alerts containing top recommendations prioritized by an ML process on the basis of efficacy, popularity, novelty and time-sensitivity. Time to market for a new recommendation, from idea to delivery, was shortened to 1-2 days, and at any given time there were up to 50 different recommendation types, rotating in and out based on their effectiveness. In the last two months of the Bank's existence, this setup was adopted to produce comprehensive reports on key bank metrics customized for each executive and delivered to them intraday via specially formatted emails.

Leading R&D function within the Data and Analytics department, providing regular quantitative analysis for the President and CEO, covering the state of the markets and KPI's of the Bank, as well as the applicable macroeconomic insights. (E.g., stability of deposits across different industries; elasticity of consumer demand manifested in outflow of funds to high-yield deposit accounts at other banks; or the impact of changes in local real estate markets on collateral valuations.) The group was also tasked with quantifying the impact of various initiatives, designing credit approval models for new loan products, analyzing client referral statistics and assisting the marketing department with NPS and lead bank analyses.

On-demand support of the C-suite and the Board with actionable analysis of the Bank fundamentals, its markets and peers; strategic initiatives and what-if scenarios. In addition, the department was responsible for facilitating internship programs at top universities, including an externship with the Computer Science department at MIT, and an Operations Research program at Princeton. The teams had broad purview over all dimensions of the Bank's operation, including deposits, loans, wealth management and operations.

Development of tools for active loan portfolio management in the context of Enterprise Risk. Early-warning system for changes in risk profile of the credit portfolio. "What-if" platform for analyzing the impact of various initiatives on capital ratios. Economic cycles model for simulating the paths of CCAR macro-indicators. Portfolio allocation tool based on derivative pricing.

One of the early quants, working on development and governance of quantitative risk models for Basel II stress testing and economic capital applications; PD, LGD and LEQ scorecards; ALLL model; wholesale credit rating migration model; Basel models for operational losses; and elements of the enterprise risk architecture design and documentation.

Founded a company to create an innovative service that enabled users to search and discover music by humming a melody and empowered artists to advertise online based on musical similarities of their work to the popular items in the catalog. Used acoustic signal processing and ML to automatically transcribe soundbites into notes and compare them to a sample catalog assembled from MIDI files and sheet music.

CashNetUSA (a FinTech startup that became part of Enova International) pioneered the use of technology to automatically process payday loan applications. My job was to develop and test probability of default, loss given default, and profitability models based on historical loan applications and performance data and to develop customer lifetime value estimates for use in the lending app and to inform the lending strategy.

Researched topology of complex networks and its application to ascribing functional roles to network nodes. Applied the findings to the study biochemical pathways in human cell biology, with molecules acting as nodes and their reactions as links. Discovered a novel approach for partitioning networks into clusters.

Developed signal processing algorithms to modify a land-based radar for use on an oil rig, using inertial navigation system to compensate for the motion of the sea, as part of the Intercontinental Ballistic Missile Defense System initiative. Designed the Kalman filter to mitigate INS errors. Briefly worked on the Patriot System design. Member of the strategic task force joint with the Office of Naval Research.

A grant-based startup performing feasibility study of a microscopic robot, propelled inside a human body by magnetic fields of MRI-like coils, to be used in targeted diagnostics and drug delivery inside the GI tract. Winner of 1999 CIMIT award (MIT/Harvard Med/Biotech). Work done in conjunction with MIT Thesis.

Theses: Dynamics and Control of Magnetic Manipulators (Robotic Surgery) Neural Network Model to Encode Location of Sound Sources in Auditory Cortex

Minor: Biomedical Engineering Thesis: Effect of Music on Studying (Attention, Retention and Cognitive Processing)