AI Operation Manager

We are seeking an AI Data Operations Manager to lead our new AI Services division. Unlike traditional BPO Operations Managers who manage established voice/chat workflows, this role requires a "Translator" and an "Architect".

Your primary mission is to take raw, often ambiguous requirements from our clients (primarily Chinese Tech & AI firms) and convert them into operational realities. You will not be building the AI; you will be building the human workforce that feeds the AI. You will define how 50-100+ agents verify various data that may be fed to client tools such as 3D radar data, correct audio translations, or annotate e-commerce images with 99% accuracy using the client's proprietary tools.

Responsibilities

1. The "Translation" (Client to Ops Bridge)

• Operationalize Vague Requirements: Digest high-level client goals (e.g., "We need better self-driving car data") and break them down into specific, trainable tasks for agents (e.g., "Annotate Lidar point clouds to distinguish between pedestrians and cyclists with <2cm margin of error").
• Guideline Localization: Take translated training manuals (often from Chinese to English) and rewrite them into "BPO-friendly" Standard Operating Procedures (SOPs) and decision trees that Filipino agents can easily understand and execute.
• Feedback Loops: Act as the primary operational point of contact for client product managers. When the client's AI model fails (e.g., confuses a dog for a cat), you must quickly update the agent guidelines to catch these specific edge cases in the next batch of work.

2. Production Management (The "Floor")

• Multi-Modal Workflow Management: Oversee diverse workstreams simultaneously, including:
• Audio Correction: Verifying automated speech-to-text transcriptions for accents and dialects.
• Computer Vision (2D/3D): Bounding box annotation for e-commerce images or 3D Point Cloud/Radar annotation for autonomous driving.
• RLHF (Reinforcement Learning from Human Feedback): Ranking AI text outputs based on safety, helpfulness, and tone.
• Capacity Planning: Manage fluctuating volumes of data. Unlike call queues, data work comes in "batches." You must effectively staff and schedule agents to clear batches before client deadlines (SLAs).
• Tooling Optimization: Master the client’s proprietary annotation tools (often web-based dashboards). Identify bugs or inefficiencies in their tools and report them back to the client's engineering team to improve agent speed.

3. Quality Assurance & Training

• "Golden Set" Management: Manage the creation of "Golden Sets" (perfectly labeled data used to test agents). Ensure your QA team is grading agents against the current client truth, which may change weekly.
• Root Cause Analysis: If a specific team is failing quality checks, determine if it is a skill issue (agent needs training), a guideline issue (the rules are unclear), or a tool issue (the software is buggy).

Specific "Translation" Examples You Will Handle

• Client Request: "Train the AI to understand English accents." -> Your Ops Process: Hire agents with specific regional dialects to record scripts and grade the AI's ability to transcribe them.
• Client Request: "Check our e-commerce search results." -> Your Ops Process: Create a workflow where agents compare a search query (e.g., "red summer dress") vs. the image result and tag it as "Relevant," "Irrelevant," or "Offensive."
• Client Request: "3D Radar Annotation." -> Your Ops Process: Train agents on spatial awareness to rotate 3D maps and draw cuboids around vehicles, ensuring no pixels are missed.

Qualifications

Experience:

• 3-5+ years in BPO Operations, preferably in Non-Voice, Content Moderation, Trust & Safety, or Data Annotation.
• Preferred experience managing campaigns with Chinese clients or Asian tech giants is a massive advantage (familiarity with the speed/culture of execution).
• Proven track record of launching new campaigns (Pioneer Accounts) where processes had to be built from scratch.

Skills:

• "Tech-Savvy" but not a Coder: You do not need to know Python. You do need to be comfortable figuring out complex, sometimes untranslated software interfaces and teaching others how to use them.
• Operational Agility: The ability to change SOPs mid-week because the client updated their data model.
• Data Literacy: Ability to look at a spreadsheet of 50,000 tasks and calculate accurate "Average Handle Time" (AHT) and "Throughput per Hour" to forecast hiring needs.

Soft Skills:

• Ambiguity Tolerance: You are comfortable working when the answer is "we are figuring it out."
• Cross-Cultural Communication: Ability to interpret "Chinglish" or direct translation instructions and grasp the intent behind the request.