Eyosias Ferid

- Lead a global TPM organization supporting 300+ engineers, researchers, and PMs across Speech, Vision, and Multimodal AI, ensuring aligned delivery on multi-year product and platform roadmaps that power billions of users. - Define long-term strategy for core recognition platforms (speech, image, multimodal) by partnering with Product, Research, Infra, and Business leadership, resulting in cohesive investment plans and reduced duplication of efforts across orgs. - Drive execution of company-critical AI initiatives (e.g., new model launches, infra migrations, cross-product integrations), consistently delivering on-time milestones for Search, YouTube, Photos, Android, and Assistant. - Champion Responsible AI and governance processes by aligning Legal, Privacy, and Trust & Safety with engineering practices, reducing policy and compliance escalations while enabling safe model deployment at scale. - Optimize cloud and compute resource usage for large-scale training and serving, collaborating with infra teams to reduce operational costs while maintaining or improving SLAs and model performance. - Mentor senior TPMs and emerging leaders through structured coaching, calibration, and succession planning, strengthening execution maturity and strategic thinking across the TPM function.

- Orchestrated end-to-end delivery of core speech and image recognition platform upgrades, coordinating multiple engineering, research, and product teams to roll out new models to flagship Google products. - Established standardized program frameworks (e.g., intake, prioritization, execution health, risk tracking) that improved roadmap predictability and increased on-time delivery rates across 10+ concurrent AI programs. - Managed the lifecycle of major AI initiatives from concept through launch, including requirements alignment, dependency mapping, and launch readiness reviews, reducing last-minute launch blockers and rework. - Implemented consistent evaluation, fairness, and safety testing workflows for speech and image models by partnering with Responsible AI and research teams, accelerating review cycles while maintaining quality thresholds. - Owned budgeting and vendor engagement for large-scale data labeling, evaluation, and research support, ensuring that external partnerships aligned with internal quality, cost, and compliance expectations. - Influenced senior engineering and product stakeholders through clear storytelling, tradeoff framing, and data-driven recommendations, enabling informed decisions on model rollout scope and timing.

- Led architecture and implementation of next-generation speech and image recognition systems using transformer-based and vision transformer (ViT) models, improving accuracy across key markets and languages. - Designed scalable training and evaluation pipelines on distributed compute infrastructure, reducing model training duration by ~50% and enabling faster experimentation cycles for research teams. - Partnered with YouTube, Photos, Assistant, and Android teams to standardize on shared recognition services and APIs, decreasing fragmentation and lowering maintenance overhead across product lines. - Improved service reliability and latency by collaborating with SRE teams to refine SLIs/SLOs, implement robust fallback mechanisms, and optimize model serving paths for real-time workloads. - Mentored Staff and Senior-level engineers through design reviews, architecture deep dives, and career guidance, raising the technical bar and improving solution quality across multiple teams.

- Developed a distributed data-processing engine using Dataflow/Beam to handle high-volume batch and streaming workloads, increasing data throughput by ~3× while lowering compute costs. - Built automated CI/CD pipelines integrated with Borg and internal tooling, reducing deployment times and enabling a shift from bi-weekly to near-daily production releases. - Implemented robust monitoring, alerting, and logging strategies by leveraging internal observability platforms, shortening incident detection and resolution times and improving operational reliability. - Collaborated with security and compliance teams to integrate access controls, encryption, and audit requirements into data pipelines, ensuring adherence to internal and regulatory standards. - Supported cross-team initiatives as a technical lead by guiding design discussions, coordinating interface contracts, and aligning delivery timelines, reducing integration friction and rework.

- Engineered microservices responsible for large-scale data ingestion and transformation using Spanner, Bigtable, and internal GCP infrastructure, enabling higher throughput and reliability for downstream systems. - Designed service APIs, data models, and schemas that supported evolving product requirements while maintaining backward compatibility and minimizing migration overhead. - Implemented dashboards and health metrics using internal monitoring tools, providing better visibility into system performance and enabling teams to proactively detect and resolve issues. - Collaborated with PMs and UX for internal developer tools by translating user needs into technical requirements, delivering features that improved usability and efficiency for internal stakeholders. - Mentored junior engineers and interns through code reviews, pair programming, and onboarding support, accelerating their ramp-up and improving code quality across the team.

- Developed backend components for indexing and batch data-processing services using Java and C++, contributing to stable and performant core infrastructure used by multiple internal teams. - Implemented unit and integration test suites that significantly increased code coverage and reduced regression bugs, improving overall confidence in releases. - Optimized existing service logic and data-access patterns, reducing latency and improving resource utilization for several high-traffic internal workflows. - Collaborated with senior engineers during design and implementation phases, incorporating feedback to learn best practices in system design, code structure, and operational readiness. - Documented system behavior, APIs, and operational runbooks, making it easier for new team members to understand and maintain services, and reducing onboarding time.