James Hosker Email & Phone Number

New York, Ny, Us

Applied AI Machine Learning Lead Data Scientist and Engineer developing ML model for fraud detection and false positive reduction as part of ML Wholesale Payments Technology (WPT) team in corporate and investment banking division of JPMC. Developed ML models using feature engineering for 3 to 10 TBs of data in Python using cloud-based platforms including Hadoop/Edison utilizing PySpark; and Amazon/Sagemaker/S3 utilizing AWS. Optimized model parameters and deployed into production ML models such as XgBoost, catboost, decision trees, RNN and others. Developed automated reports indication performance of models (tables and plots) based on recall, precision, Shapley graphs, ROC-AUC, score distributions, model feature of importance, false positive reduction ratios, TPs vs. FNs, coverage ratio metrics and many others. Using several python libraries including PySpark, Tensorflow, Keras, AWS, Seaborn (sns), SciKit learn, Matplotlib, SciPy, and many others.

Head of mortgage quantitative analytics for modeling, PnL and risk of actively managed mortgage and UST Exchange Traded Funds (ETF), working directly with portfolio managers. Extensive coding of analytics, models and backtests in Python & Microsoft SQL using Spyder as well as both Yield Book (YB) & BAM/PORT models. Largest ETF manages a $6 billion portfolio in mortgages including agency pass-throughs, CMO, CMBS, ABS, TBA, RPL, NPL, UST, bond futures, and non-agency RMBS. Utilize machine learning methods for some analytics (sklearn, keras).

Paris, Fr

Recruited to worked with hedge funds, investment funds, pension funds and insurance companies in designing macro cross-asset alpha and hedging strategies. Strategies and products utilized equity, credit, FX, rate and commodity derivatives; ETF options; VIX futures; and VIX options. Organized a team that prices, back tested and generated transaction on hedging strategies and smart-beta strategies for clients that generated €20 MM/yr. in sales revenues. Coded in Python, SQL and VB.

New York, Ny, Us

Recruited to build out a sales strategy for an electronic corporate bond trading platform. Company funded by Warburg Pincus as a start-up in disruptive financial technology. Manage a team and built real-time alerts and relative value screens. Coded in SQL, Scala & C++.

Paris, Fr

Recruited to create macro cross-asset trading strategies for hedge funds and other clients using credit, equity, FX, commodity and/or rate derivative hybrids. Successfully developed index, sovereign and individual CDS and bond trade ideas, which generated €15 MM/yr. in sales revenues. Built a credit index volatility infrastructure for swaptions business and structured first-to-defaults (FTD) options; credit linked notes (CLNs); and other bespoke products. Supervised a team. Code in C++, SQL and VB.

New York, New York, Us

Developed platform to measure equity analyst stock picking accuracy and earnings forecasts for the conviction list. Appointed as member of the Investment Review Committee (IRC). Supervised team, built screens, and back tested thematic ideas using fundamental data (e.g. balance sheets and cash flow data for M&A targets). Coded in Slang/SecDb.

New York, New York, Us

Head Strategist for Capital Structure Franchise TradingRecruited to develop trade ideas, hedges, and pricing models for real-time risk and PnL platform for the distressed equity and debt trading desk, which generated $250 MM/yr. in 2006 & 2007. Implemented proprietary strategies using credit-equity and volatility relationships, generating $20 MM/yr. Team traded converts, CDS, bonds, stock, warrants, preferred stock and equity options on the same companies. Developed “Fall98” scenario to perform mark-to-market (MTM) shocks across assets, implemented hedges that prevented losses in 2007-2008, and supervised team of strategists. Coded in Slang/SecDb.

Us

Responsible for the development, marketing, publication and global coordination of derivative research for equity option volatility and quantitative models. Supervised and managed a team of four individuals which ]ranked #2 in 2005 and #3 in 2004 for both the Institutional Investor (II) and Hedge fund surveys in Equity Derivatives Research. Built monitors on index dispersion trades. Back tested and traded index rebalances (2001-2004) using TAC data, generating $15 MM/yr. Coded in C and VB.

Us

Built and managed Latin American equity derivatives flow and structured business (generated $25 MM/yr). Exotic options trader for baskets, quantos, barriers, index options, collars, notes, etc. (generated $20 MM/yr). Listed options trader for biotech and tech stocks, including buy-writes and overwrites (generated $20 MM/yr).

Mclean, Va, Us

Project leader responsible for the design, test and delivery a multi-media system to display sensor imagery from planes and satellites used in Operation Desert Storm. Presented to high ranking US and NATO military officials. Built complex analytical models to assess image quality for an FBI recognition system. MITRE is a Federally Funded Research & Development Center (FFRDC). Awards: MITRE Graduate Degree Award and MITRE Program Achievement Award from CEO. Coded in Pascal, C and Adaptec.

Mba, Finance

Master Of Science (M.S.), Data Science

Msee, Electical Engineering

Bsee, Electrical Engineering