Development

Surgical Video Annotation Program Lead

Pune, Maharashtra
Work Type: Full Time

Job Title: Surgical Video Annotation Program Lead (Team Lead) 

 

Location: India - Pune 

Type: Full-time 

Reports to: Head of Delivery / Program Director 


About Us

At Codvo, we are committed to building scalable, future-ready data platforms that power business impact. We believe in a culture of innovation, collaboration, and growth, where engineers can experiment, learn, and thrive. Join us to be part of a team that solves complex data challenges with creativity and cutting-edge technology.


Role goal: Own end-to-end delivery of a high-throughput, clinically defensible, audit-ready surgical video annotation program—driving automation-first workflows, quality (IRR/QA), and on-time dataset releases.

 

What you will own 


Program delivery (E2E): Stand up and run the annotation “factory” from intake → de-ID → task orchestration → annotation → QA/IRR → adjudication → dataset release + evidence pack.


Ontology + guidelines execution: Partner with clinical SMEs to operationalize a procedure-specific ontology (phases/steps, tools, anatomy, events) and convert it into clear labeling guidelines and UI rules.


Automation-first operations: Drive pre-labeling + verification workflows (not manual-from-scratch), implement routing based on model confidence/uncertainty, and continuously reduce human effort per labeled minute.


Quality system: Implement and enforce:


Multi-rater sampling strategy


IRR reporting by label type (kappa/alpha; IoU/Dice where applicable)


Calibration loops and retraining for annotators


QA gates + sampling plans with acceptance thresholds


Adjudication governance: Run the disagreement workflow, manage escalation to senior annotators/clinical reviewers, track ambiguity categories, and ensure guideline updates close recurring issues.


Dataset release management: Own versioning, provenance, and release discipline—ensuring every dataset is reproducible and ships with an audit-ready Evidence Pack (provenance, QA, IRR, adjudication trail, sign-offs).


Security + compliance coordination: Ensure labeling operations follow enterprise security requirements (access control, logging, retention, de-identification review) and support audits/vendor risk requests.


Client-facing cadence: Lead weekly operating reviews, present throughput/quality metrics, manage scope changes, and ensure PoCs convert into scaled programs.



What you will build and run 


Team: L1 annotators, L2 senior annotators, QA auditors, adjudicators; coordinate with clinical reviewers and ML/data engineering.


Operating system: SOPs, training curriculum, calibration playbooks, quality scorecards, escalation paths, and release checklists.


Metrics: Throughput, cycle time, rework rate, IRR trends, defect density, acceptance pass rate, cost per labeled hour/minute, and automation leverage (pre-label acceptance rate).



Required qualifications 


6–10+ years in annotation operations / data operations / QA-led delivery, with at least 2+ years in a lead role managing teams and SLAs.


Hands-on experience with video annotation (temporal segmentation + event labeling) and familiarity with bounding boxes/segmentation concepts.


Demonstrated ability to implement multi-rater workflows, compute/interpret IRR, and run calibration to improve consistency.


Strong program management skills: planning, staffing, throughput modeling, risk management, and stakeholder communication.


Comfort working with tooling/APIs and structured data exports; ability to translate guidelines into tool-enforceable rules.


Experience in regulated or sensitive-data environments (healthcare preferred): privacy-first mindset, audit trails, process discipline.


 


Preferred qualifications (strong plus) 


Healthcare domain familiarity: surgical workflows, OR video sources (endoscopy/robotic), common quality issues (smoke, blur, blood occlusion).


Experience coordinating de-identification workflows for video/audio and supporting enterprise security reviews (SOC2/ISO-type controls).


Exposure to automation/ML-assisted labeling: pre-labeling, confidence routing, active learning basics.


Prior work on dataset versioning and “release” discipline (e.g., DVC-like thinking, evidence packs, reproducible builds).


Note- Please apply via our official careers portal only, as applications sent directly to executives may not be considered. 

Submit Your Application

You have successfully applied
  • You have errors in applying