Senior Program Manager, Agentic Ai
https://docs.google.com/document/d/1A2oThuc5g1flKVraq-N9i2GCnVdmkmz336ZUMs3ASrA/edit?tab=t.0 ~~ ~~ ~~
About the role and team
The Uber Localization Team is seeking a Senior Operations Program Manager
(Multilingual GenAI PgM) to lead the evolution of our multilingual operations through Agentic
workspace using the Large Language Models (LLMs).
In this role, you will designing and building the end-to-end ML strategy for the Localization
operation team, with a focus on selecting and experimenting the best LLM model for
integrating into production workflows. You will evaluate, select, and operationalize LLM
solutions (including hybrid approaches) across high-resource languages (HRLs), low-resource
languages (LRLs), and low resources content types, leveraging Google Cloud Platform (GCP)
services (e.g., Vertex AI, AutoML Translation, BigQuery) or any cloud providers to balance
quality, scalability, latency, and cost.
You will partner closely with Localization, Engineering, Product, and MLOps/LLMOps teams to
translate experimental capabilities into production-grade systems, while ensuring
measurable impact on translation quality, turnaround time, and cost efficiency
What you’ll do
ML/GenAI Strategy & Evaluation
- Define model selection strategy by use case (HRL vs LRL, real-time vs offline,
high-risk vs low-risk content). - Run experiments (prompting, fine-tuning, RAG, glossary constraints) and translate
findings into production recommendations. - Select and evaluate LLM models or hybrid approaches using structured frameworks
(e.g., BLEU, COMET, human eval, task success metrics). Evaluate and benchmark GCP-native models and services alongside external
vendors, making recommendations based on quality, latency, and cost
Operational Pipeline Ownership
- Design and operate localization pipelines and leveraging GCP services (e.g., Vertex AI
for model hosting/evaluation, AutoML Translation, TLLM, Cloud Functions,
BigQuery) to enable scalable, production-grade LLM workflows. - Build and optimize pipelines for throughput, latency, and cost efficiency, including
fallback strategies. Drive automation (scripting, APIs, orchestration) to reduce manual effort and operational
overhead.
Productionization & Cross-functional Execution
- Define requirements and partner with Engineering to productionize LLM solutions.
- Partner with Engineering and MLOps to deploy and monitor models on GCP
infrastructure, ensuring reliability, observability, and cost control. - Write clear PRDs and operational specs covering data flow, model behavior,
evaluation, and guardrails. - Work with MLOps/LLMOps to ensure monitoring, versioning, and lifecycle
management of models.
Quality & Performance Management
- Establish quality frameworks combining automated metrics and human evaluation.
- Monitor model performance in production and drive continuous improvement loops.
- Own trade-off decisions across quality, cost, and speed, with clear reporting to
stakeholders.
Stakeholder & Program Leadership
- Act as the primary point of contact for ML-driven solutions across internal teams and
vendors. - Align stakeholders on roadmap, priorities, and measurable outcomes.
Lead cross-functional initiatives across time zones.
Data & Insights
- Partner with Data/Engineering to build dashboards tracking:
- Model performance (quality metrics)
- Cost per word/request
- Throughput and latency
- Enable data-driven decision-making for both technical and business stakeholders.
Basic Qualifications
Bachelor’s degree in Computer Science, Engineering, Linguistics, or a related field; or 4 years of equivalent professional experience.
- 3+ years working with MT, LLM, or ML-driven language systems in production
including deployment. - 1+ years in localization operations (CMS/TMS) in a tech or platform environment.
- Proven experience driving end-to-end programs, not just execution within a single
function.
Preferred Qualifications
3+ years of experience leading large-scale MT/LLM transformation programs.
Hands-on experience fine-tuning LLMs for multilingual use cases or implementing RAG pipelines for grounded translation.
Proficiency in Python scripting or API integration to automate data flows and reduce operational overhead.
Demonstrated ability to build data dashboards that track model performance, throughput, and cost-per-word metrics.
Proven track record of navigating high-ambiguity environments and building frameworks for unique technical challenges.
Responsibilities
For Sunnyvale, CA-based roles: The base salary range for this role is USD $167,000 per year - USD $185,000 per year.
You will be eligible to participate in Uber's bonus program, and may be offered an equity award & other types of comp. All full-time employees are eligible to participate in a 401(k) plan. You will also be eligible for various benefits.
Ready to Ride?
This isn't the kind of place where you follow a playbook — it's where you help write one. If you're driven by impact, energized by challenge, and ready to shape how the world moves — we'd love to hear from you.
You may be eligible for bonuses, equity, and other compensation, as well as a range of benefits. Explore our benefits.
Offices remain key to collaboration and Uber's culture. Unless approved for full remote work, employees must spend at least 50% of their time in-office. Some roles, like those at greenlight hubs, require full-time in-office presence. Ask your Recruiter for details about this role's requirements.
Uber is proud to be an Equal Opportunity employer. All qualified applicants will receive consideration for employment without regard to sex, gender identity, sexual orientation, race, color, religion, national origin, disability, protected Veteran status, age, or any other characteristic protected by law. We also consider qualified applicants regardless of criminal histories, consistent with legal requirements. If you have a disability or special need that requires accommodation, please let us know by completing this form.