Abraham Howland

 ahowland@gmail.com |  | 

(June 2023 - Jan 2025)

As the sole Principal Software Engineer in the Enterprise Core Operations (ECO) space, I had two primary responsibilities:

• to coordinate the efforts of the several ECO engineering teams that managed services exposing core data (through the Core Services Program), and
• to oversee the re-architecture of a unified GraphQL API view into core data (Wentworth)

Core Services Program

The Core Services Program's goal was to retire legacy mainframe systems, replacing them with modernized, distributed components organized by domain. My work involved defining requirements, interface boundaries, and shared architectural patterns outlining the expected technical capabilities of core systems, in order to ensure that all other systems at H-E-B could access core data in a unified and controlled manner.

Wentworth

Wentworth was a GraphQL API offering a unified view into H-E-B's core data. My role involved auditing, documenting, and re-architecting Wentworth's sub-systems, as well as resetting the team's contracts and working relationships with its peer (core data) teams.

(May 2021 - June 2023)

Pricing Core Service (PCS)

As the technical lead on the PCS team, I built micro-services that managed core pricing data along three dimensions: UPC, location, and time. I also built backend services that enabled H-E-B to flexibly apply different e-commerce markup values in fine-grained ways to (arbitrary) sets of products at (arbitrary) sets of locations over (arbitrary) intervals of time. Our services, jobs, and APIs were written in Haskell and Java, interacted with Postgres and Oracle databases, were deployed using Kubernetes, and communicated internally and externally via RabbitMQ, Pub/Sub, TIBCO, and HTTP.

Frontstage

One of the most pervasive problems I faced at H-E-B involved engineers' inability to gain awareness into data dependencies between internally managed software systems. To tackle this problem, I built Frontstage, a backend service and frontend application for reifying and visualizing graph representations of the systems owned by core teams, and their relationships to other systems. I prototyped Frontstage in Haskell and Elm, and then transferred ownership of the application to an engineer who rewrote the backend in Java, and leveraged AngularJS in the frontend.

(March 2018 - April 2021)

Flipstone is a small group of consultants / contractors, and is responsible for bringing Haskell to H-E-B. I taught Haskell to a team of Java developers at H-E-B, and was embedded on two core teams (CHEDR and WIMS), as part of an effort to build resilient software that would aid H-E-B's mainframe retirement efforts.

Cost Core Service (CHEDR)

I created APIs for managing the cost of (cases of) items at (warehouse) locations, indexed by time. I also wrote a service that pulled external commodity cost data from Bloomberg.

Purchase Order Core Service (WIMS)

I built a core service that managed purchase order data, replacing behavior previously performed by the purchasing component on the mainframe (an extension of OMI's BICEPS software). The service was written in Haskell, backed by a Postgres database, and deployed to GCP. Integrations with TIBCO EMS enabled sending and receiving EDI messages representing updates to the statuses of purchase orders.

Libraries (internal to H-E-B)

(September 2016 - March 2018)

As independent contractor, I helped a client (ITPro.TV) that provided educational video content to subscribers by:

• rewriting portions of a Node backend into Golang, and
• using Terraform to improve the stability of deployments

Haskell | Elm | Purescript | Python | Java | Golang | Javascript | Bash | Git | Gitlab CI/CD | Docker | Kubernetes | GCP | Datadog | Terraform | Linux | Postgres | Oracle | Apache AGE | RabbitMQ | Kafka | TIBCO | Pub/Sub | IBM MQ

Category Theory | (History of) Philosophy | (History of) Computing