Benjamin Manes Email & Phone Number

San Francisco, Ca, Us

At Vector, we enable contactless pickup and delivery through collaborative workflows for shippers, receivers, and carriers.

San Francisco, Ca, Us

• Developed portals to allow carriers, brokers, and customers to interact throughout the life-cycle of an order.• Real-time integration with AS/400-based logistics product. This enables the largest fleets and brokers to migrate off of their legacy system without impacting other business workflows during the transition (such as accounting and payroll).• Enhancing the processing engine to support long running, transactional workflows that span multiple entities.• Created a generic data import process by converting csv to json using the schema definitions. Then wrote simple mappers for customer files to our internal format, which are transformed and batch processed for import.• Designed a metadata-driven architecture, workflow engine, and co-designed the UI component model to allow for rich customization and a faster development process. Reimplemented the server-side and bridged with the legacy application by using RabbitMQ events to synchronize during the migration.• Rewrote the server from Django to Java to provide a baseline before rearchitecting the product.• Redesigned the Android document capture process, introducing a lifecycle that resolved the camera crash and other quality issues. Proposed an algorithm that stabilized edge detection by taking the frequency and recency of candidate squares within a sample period.• Added instant search with real-time suggestions, replacing a slower SQL-based full text search implementation.• Implemented persistent task queues and offline synchronization for Android and iOS applications.• Numerous tasks related to development of the web (React), server (Python), and mobile (ReactNative) applications.

New York, New York, Us

• Led the platform team.• Numerous projects to improve performance, pay down technical debt, and implement core services.• Replacing a home-grown MySql abstraction that stores Java binary objects and performs class migrations with relational modeling, jOOQ, and Flyway migrations.• Transitioning the platform towards a lightweight service-oriented architecture for application development, restful endpoints for externalization, map/reduce for data processing workflows, and an event stream to decouple layered servers and services.• Introduced Continuous Delivery to increase reliability by minimizing disruptive changes due to large releases, manual testing, and complex merging strategies.• Implemented a cache-coherent distributed graph storage layer designed for frequent traversals and ACID characteristics. The design utilizes persistent data structures and transaction logs to allow for a high degree of caching and remaining fault tolerant.• Led a project to achieve stateless servers to scale an architecture originally designed with support for only a single production machine.• Replaced an ad hoc build system with Gradle (with Maven as a stopgap), standardized the coding style and module conventions, introduced modern testing frameworks, presented on Google Guice best practices, laid out roadmap for scaling, etc.• Provided guidance for the design of REST apis, bootstrapped the team with a server written in Scala, and implemented core routes.

Mountain View, Ca, Us

• Participated in an urgent project with executive-level visibility to streamline the sales process.• Due to the success of Patchwork, the team has grown from 3 engineers to 5 (plus 1 intern). The project's scope has increased to define an open platform for enterprise applications at Google.• Maturing the back-end architecture of Patchwork based on a more flexible data model. It supports configurable and personalizable applications, role-based views, ACLs, browser synchronization, and shared components.• Invited to speak at the Strange Loop conference on caching and lock amortized data structures.• Received five peer bonuses and two management spot bonuses.

Mountain View, Ca, Us

Adding support for a maximum capacity and an eviction listener to Google Guava's MapMaker. This work is based on my open source project.Co-designed the architecture and acted as a core developer for Patchwork: a Gwt application framework with native OpenSocial gadget support. Specifically I was involved in...• An agent-based design for layout handling and gadget rendering based on the Presentation-Abstraction-Control pattern.• The usage of data overlays to support user customizations, sharing of user-defined templates, and role-based views.• A non-blocking user experience through soft client-side state, eventually consistent storage, and a write conflict detection algorithm.• The performance and reliability monitoring, and operational readiness for releasing to production.

Oakland, California, Us

• Led weekly architecture discussion group and member of Platform Software Architecture Review Counsel (PSARC).• Redesigned the company and configuration management models as replacements to legacy rules engine, cloning, and replication processes. The design focuses on scalability and leverages an event-driven architecture with a well defined data life-cycle management process.• In coordination with other team members, developed platform services for configuration and user profile management. Aided in the migration from legacy APIs to their replacement services.• Developed a metadata-driven caching framework that supports local and remote caching layers, disk caching, indexed lookup, expiration policies, and integration with internal Spring-based frameworks.• Implemented a rich utility layer for a green-field platform to promote a bottom-up development style based on the Google Collections open-source project.

Oakland, California, Us

• Recognized as a “Rearden Rockstar” for infrastructural improvements that delivered significant performance gains and a dramatically improved customer experience.• Implemented bulk user profile management capabilities, such as for performing large scale extractions and uploads by administrators.• Designed a distributed master-worker framework for processing large data sets across a cluster of nodes. Proof of concept fork-join and map-reduce abstractions were prototyped, but were not matured to production quality.• Added the capability for administrators and technical support staff to perform actions on a user’s behalf through an audited impersonation flow.• Developed a new user interface and data mining capabilities for an internal management dashboard used by the Operations team to quickly identify and resolve production incidents.

M.S., Computer Engineering

B.S., Computer Engineering

B.S., Computer Science