Michael Comes Email & Phone Number

San Francisco, California, Us

San Francisco, California, Us

Integrated 3rd party APIs for supporting marketplace lending, underwriting, payment transmission, and enhancing financial instruments platform capabilities. Synchronize payment system funding requests with application workflows, tying interest rate calculation mechanisms to real-time fund movements, replacing heuristic-based approaches. Designed and prototyped durable ledger processing system Implemented payment transmission and cancelation components of investor payments ledgering and transmission system for scaling the implementation of DACA (Deposit Account Control Agreement) accounting system.Develop secure webhooks ingestion API, including, models, mappers and obfuscation to prevent security breaches. Designed and implemented graph of finite state machine progression for underwriting and financial product outcomes for historical state reconstruction. Built as a callable API to enable job-based backfills.Refactor system architecture to support 3rd party underwriter decisioning with minimal impact on existing onboarding flows, including full audit trails. Built business logic to manage fractional pennies and negative balances for precise accounting of payment transaction ownership b/w investors for financial operations.

San Francisco, California, Us

San Francisco, Ca, Us

Design and implement back-end data services and ETL/data-processing pipeliness and jobs on the data platform team2019 Womply Hackathon 1st place: Built a Developer Portal and API platform utilizing Apigee, allowing 3rd party clients to consume small business benchmarking data API. Included configuration of token-based authentication, CORs policies (allowing for embedding API calls into UI components), and exposing endpoints behind VPC firewalls to public internet. Utilized as prototype for Q3 Developer Platform roll-out. Implemented back-end service integration with Plaid, allowing CRM software users to upload bank statements from Plaid and receive notification via webhooks when bank loans were available from 3rd party cash flow lenders.Build stochastic email-sending functionality for automated email marketing product, allowing internal developers to fine-tune email-blocking thresholds based on domain and mx records, for large number of email send requests.Optimization and fine-tuning of hosted kafka cluster for social data system, which scrapes millions of social media pages to support CRUD operations of social media reviews by CRM users in web app. This involved redesign of offset storage system by migration from in-memory and zookeeper, and kafka, to kafka-only, solving latency problems when consuming messages. Resulted in higher message throughput and better performance in CRM web-app.Design and implement self-service portal web API by triggering a real-time social media page scrape to populate app dashboard once a user has selected to start a free trial. Solves the zero-data problem by giving business owner immediate access to relevant application data once a free-trial has started. Resulted in 20% increase in sales efficiency.

Jacksonville, Florida, Us

Lead 3-person dev team for migration from z/os architecture to containerized, cloud-based nodeJs micro-services. Leverage Kubernetes and docker platforms for continuous health-checks and seamless redeployment and pod autoscaling during periods of CPU load, roughly 300k requests / hour during peak usage.Develop multi-threaded java batch process to verify 2.2 million email addresses in less than 45 minutes.Write Java back-end REST services utilizing Websphere, Spring framework, SpringBoot for rapid deployment. Configure services with build pipelines utilizing Jenkins2, Gradle and Kubernetes onto Openshift platform. Develop cache solutions in Redux stores to reduce component rendering times to 75ms from 300 ms.Implement and architect React-redux front-ends for web applications and migration from existing .jsp infrastructure. Utilize native React-router library and create split-points between container routes to minimize load- and rendering-times. Cache large code dependencies utilizing CommonsChunk plugin across multiple app entry points. Implement multi-threaded, AsyncWorkManager library in Java to call NodeJs service implementation for website look-up of key data. Built and deployed scalable NodeJs and Java APIs to interface NodeJs services with existing messaging infrastructure. NodeJs library to be published on firm-wide image repository. Can handle usage of more than 1000 users, 4k requests per second without dropped messages or memory leaks, utilizing NodeJs event-driven model with messaging APIImplement NodeJs read and write micro-services for storage and utilization of relational and noSQL, customer interaction data. Services to be utilized for firm-wide machine-learning. Optimized to handle 8 to 12 million requests per day during open-enrollment periods. Implement role-based authorization and verification mechanism into micro-service middleware, utilizing JWT distributed infrastructure to secure confidential healthcare data.

Sarasota, Fl, Us

Developed and wrote code for Cumberland's automated, proprietary credit-monitoring software. Software assigned scores to individual bonds based on an issuer’s financial and industry metrics, utilizing the Bloomberg API and 3rd party database. Used as an alternative to hiring additional staff; electronically monitoring financial data releases for 4800 CUSIPs on real-time basis; resulted in procurement of additional $1 billion from client. Now used with regional banks to meet Dodd-Frank regs; pitched as an outsourced credit product to regional banks via an information services companyDesigned and maintained database of 4800 CUSIPs from ETL to data analytics processing. Proficient in Microsoft server-side stack: T-SQL, MDX, DMX, DAX data processing and analytics languages.Wrote code in C# for http requests and web service control flow; aid in control and data flow from Bloomberg API to Excel spreadsheets and client-side GUIs, deployed to analysts and PMs on staff.Data analytics and model-building: Developed spread and rating prediction models for state GO credits to evaluate richness/cheapness of the market, utilizing ordered logit and regression analysis. Prediction accuracy of 80% within 1 notch for ratings models and R2 of 0.6 for spread prediction models. Sourced data from government statistics bureaus, Bloomberg utilizing web APIs. Implemented new-issue and secondary market trading strategy, realized gains of $200mm+ on portfolio of $2bn.Author: Adventures in Muniland: A Guide to Municipal Bond Investing in the Post-Crisis Era, Amazon Top-Seller, 3 weeks; reviewed by Barron’s, Business Insider, The Bond Buyer Magazine; Utilized as textbook at Baylor UniversityMarket Expert: Print/tv/radio: Bloomberg, Wall Street Journal, Forbes, CBS News, NPR, Barron's, Fox Business

New York, New York, Us

Developed financial models and drafted research reports for current and potential investments; wrote software trading strategies for firm's long/short equity hedge fund; assisted with database development for back-testing of software stock-trading strategies