Main technologies: backend API on Python, front end web application on React, all hosted in Google Cloud

Helped grow the number of loans Valon services from 50 to 50,000.

enabled Valon to offer forbearance plans, repayment plans, payment deferrals, and loan modifications for borrowers who are experiencing difficulties paying their mortgage (over 250 plans plans so far)

integrated with our proprietary payments ledger in order to automatically allocate payments to the plans and permanently modify the loan terms

automated onboarding of previously existing plans from prior servicers to allow Valon to compliantly service these loans (onboarded ~400 plans with an expected load of 5,000 in 2023)

integrated with Nice API (an automated dialer) to call delinquent borrowers dependent on a set of eligibility criteria, scaling the number of collections outreach from 300 to ~4k per month

created a page in the internal console to consolidate relevant delinquency information for expedited customer support

integrated with a third party vendor to automatically determine when to order a property inspection and execute that order and ingest the results via SFTP

Enabled Proxy to bring in revenue of >$1M/year by integrating with Twilio billing system

Built scalable APIs to be able to handle hundreds of millions of requests a day (data validation, read/write from MySQL DB, logging, metrics, caching)

Automated Calls Per Second billing for Twilio’s Voice product, reducing manual operation time by hundreds of total hours

Implemented Quota API (a Scala + Akka HTTP API) to empower Twilio to bill in a completely different way: quota based instead of event based – allowed sellers to make million dollar deals with the Twilio Flex product

Tools used: Python, MySQL, REST APIs, Spark, AWS, S3