Overview
Product Software Engineer II on the Metadata Authority team in Disney Entertainment and ESPN Product & Technology, building multimodal and Gen AI pipelines that turn raw video, artwork and text into structured, authoritative metadata.
What you'll do
- Build and maintain multimodal AI pipelines for extracting metadata from video, artwork, and text, including captions, object detection, scene classification, summarization, and sentiment analysis.
- Develop low-latency inference services to process live content without blocking downstream workflows.
- Integrate Gen AI capabilities into production systems using AWS (SageMaker, Lambda, S3, Glue) and distributed processing frameworks where applicable.
- Collaborate with data platform teams to consume curated data from the organization's medallion-architecture pipelines and return enriched metadata for Gold-layer publishing.
- Implement quality control, validation, and governance processes for AI-generated metadata to ensure accuracy, compliance, and auditability.
- Automate workflows and CI/CD pipelines for model training, deployment, and monitoring.
- Drive observability, alerting, and operational excellence for AI services at scale.
What you'll need
- 3+ years of backend or ML engineering experience, with strong proficiency in Python and/or Java.
- Proven track record deploying AI/ML models into production, with emphasis on multimodal and Gen AI systems.
- Proficient knowledge of AWS services (SageMaker, Lambda, S3, SQS, Glue) and experience with real-time and batch processing.
- Understanding of ML lifecycle, including inference optimization, monitoring, and model versioning.
- Experience integrating AI services into production APIs and content workflows.
- Proficiency with relational and NoSQL databases for metadata storage and retrieval.
- Bachelor's degree in Computer Science, Information Systems, Software, Electrical or Electronics Engineering, or comparable field of study, and/or equivalent work experience.
- Minimum of 3 years related work experience.
Nice to have
- Experience with vision-language models (e.g., CLIP, BLIP-2), LLMs, or speech-to-text systems.
- Knowledge of metadata classification, entity resolution, or knowledge graph enrichment.
- Familiarity with Elasticsearch or search/indexing systems.
- Background in sports, entertainment, or media content pipelines.
Details
- Full-time employment.
- Location: Bangalore, India.
Read the full description and apply on the company’s own careers page.