Atilano Garcia

AI-Powered Screening Launch: Led the production rollout of AI-generated candidate screening, enabling automated screening questions and credential checks across the hiring workflow.

• Integrated backend services, frontend application flows, scoring pipelines, feature-flagged rollout, and observability tooling to support a safe end-to-end launch.
• Improved launch safety and debugging by adding rollout controls and structured logging for new AI-driven flows.

Premium Recruiting Product Launch: Drove launch work for premium recruiting features by aligning entitlement logic, frontend flows, and product instrumentation, contributing to a product expansion tied to $6M in incremental committed revenue.

AI Screening Reliability Improvements: Improved the reliability of AI-powered screening and scoring systems by preserving user edits, separating automated and manual data paths, and mitigating model rate limits through platform-level traffic and capacity controls.

Litigation Hold Data Capture & Processing: Spearheaded the implementation of data capture for non-email communication channels, fulfilling critical legal requests for user communications and decreasing overhead by 2000%.

• Configured AWS infrastructure and IAM using Terraform, including an AWS Firehose pipeline that ingests messages from an SNS topic and dynamically separates communications into prefixes by date and type for streamlined data management.
• Used Spark to process and store millions of communications in AWS S3 tables for later indexing.
• Implemented retrieval and storage for 5 communication mediums (Email, SMS, Native Mobile Push Notifications, WhatsApp, Line, Messenger).
• Researched bloom filters and z-ordering to optimize what would become a multi-petabyte database.

Messaging Operational Improvements: Investigated, experimented with, and implemented deployment pipeline changes that resulted in a 37% runtime improvement across 5 separate code repositories, enhancing CI/CD efficiency and reducing time to deploy.

Notification Deduplication: Implemented core deduplication logic and a robust key-generation function to prevent duplicate notifications from being sent to end users.

• Integrated deduplication across multiple channels including email, SMS, in-application notifications, and browser push.
• Refactored the deduplication information service to improve understandability and maintainability.

Asynchronous GraphQL Refactoring: Converted synchronous GraphQL mutations and dataloaders to operate asynchronously, improving system responsiveness and efficiency.

• Resolved asynchronous unit-test failures across multiple critical test suites.
• Removed caching from dataloaders and documented how to reintroduce it safely in the future.

Functional Test & System Reliability

• Optimized functional tests to complete in 1 minute 6 seconds, a 96% reduction in suite run time.
• Implemented Datadog performance monitoring, analysis, and rollback procedures for the call-to-interview service and mosaic provider deploy pipelines, improving deployment safety and reliability.