Job Title

WORK MODE: Full-Time | Hybrid - Bangalore | 3 Days WFO EXPERIENCE: 5+ Years (shipping deep learning models into production)JOB DESCRIPTION:The RoleWe are seeking a Lead Deep Learning Engineer to lead the development of the company’s Visual AI stack.This is a highly technical Senior Individual Contributor role with strategic ownership. You will work with large-scale, real-world culinary vision data captured from robots operating in homes and convert that data into production-grade perception and autonomy systems.This role requires model inventors and AI architects, not just pipeline optimizers.Key Responsibilities1. Model Architecture & Native AI DevelopmentDesign and train transformer-based computer vision models from first principlesDevelop capabilities across: Segmentation | Object detection | Classification | RegressionWork extensively with Vision Transformer architectures such as: ViT | Swin Transformer | MViT | SegFormerMake principled tradeoffs across latency, reliability, cost, and deployed performanceThis role demands hands-on experience building and experimenting with transformer architectures—not merely fine-tuning pre-trained models or integrating APIs 2. Autonomy & Model StrategyDefine the appropriate autonomy targets given real-world kitchen variabilityTranslate autonomy goals into a clear Operational Design DomainDecide when to use large general models vs. distilled or task-specific modelsMake strategic shifts based on real-world constraints and deployment learningsThis is not just execution, you will influence core technical direction.3. Data Strategy & Failure-Driven LearningDefine what data should be collected and in what sequenceBalance on-device data, human demonstrations, and curated datasetsDesign robust feedback loops using: Intervention data | Edge case logging | Replay and prioritization frameworksContinuously improve reliability in deployed consumer environments4. Cross-Functional LeadershipYou will collaborate closely with:Data & Annotation teams working on culinary workflowsCore Software teams integrating perception models into autonomous cooking systemsYou will lead a highly capable team while remaining deeply hands-on.Ideal Candidate ProfileRequired5+ years of experience shipping deep learning models into productionDeep, hands-on expertise in Vision TransformersProven experience designing and training models from scratchStrong architectural intuition and first-principles thinkingDemonstrated ownership over technical decisions and model direction PreferredBroad computer vision experience across multiple perception domainsExperience with real-world, noisy, physical-environment dataBackground in segmentation, detection, and state-change recognition tasksComfort operating in ambiguity and defining strategyImportant ClarificationThis is not primarily:An ML infrastructure optimization roleA DeepStream/TensorRT-heavy deployment engineering positionA YOLO-integration or API-centric ML roleA RAG/chatbot-focused AI positionWe are specifically looking for engineers who design and invent AI models, not those who primarily scale or deploy them.Why This OpportunityWork on real-world autonomy at consumer scaleDirect impact on AI operating in physical environmentsHigh ownership and strategic influenceJoin a mission-driven team building category-defining robotics productsAbout the Company:A well-funded, Series A consumer robotics company building autonomous kitchen robots that cook complete meals in real homes. Their AI-powered systems are deployed at scale in the U.S. market and operate daily in dynamic, real-world kitchen environments.This is one of the few teams globally shipping cutting-edge AI into the physical world at consumer scale.