Benefits:
- 401(k)
- 401(k) matching
- Competitive salary
- Dental insurance
- Health insurance
- Opportunity for advancement
- Paid time off
- Relocation bonus
- Vision insurance
- Wellness resources
Petlibro | San Jose, CA
About Petlibro
Petlibro is a design-thinking company creating products that nurture the intertwined lives of pets and their people. We launched with a philosophy that good design, in form and in function, can make a difference. Petlibro innovates with the latest technology to solve everyday problems for modern pet parents and revolutionize how we care for our pets.
Since 2019, Petlibro has grown into one of the best-selling pet tech brands globally. From smart feeders with app insights to ultra-filtered automatic fountains to pet-health-focused smart apps, our products are engineered to magnify the bond between your pet and you. We are now building AI-powered products and services for modern pet care.
Job Summary
We are looking for an AI Engineer to design, train, and deploy AI models that power intelligent pet care and animal monitoring products. You will work across the full model lifecycle, from data curation and fine-tuning to production serving and on-device deployment. The role spans computer vision, vision-language models (VLMs), large language models (LLMs), and agentic AI systems, with a strong focus on animal behavior understanding and multimodal reasoning.
Key Responsibilities
- Fine-tune LLMs and VLMs for pet care and animal behavior domains using modern post-training methods (SFT, DPO, GRPO/DAPO)
- Develop and evaluate computer vision models for animal detection, pose estimation, activity recognition, and health monitoring
- Build multimodal AI pipelines that combine video, audio, weight sensor data, and other sensor modalities for real-time animal behavior analysis
- Design and operate model serving infrastructure with high throughput and low latency (vLLM, TensorRT-LLM)
- Implement RAG systems and agentic workflows for domain-specific knowledge retrieval and automated decision-making
- Optimize models for on-device / edge deployment using quantization (GPTQ, AWQ, INT4/INT8), distillation, etc.
- Curate and manage training datasets including synthetic data generation and annotation pipelines for animal imagery and video
- Develop evaluation frameworks and benchmarks specific to animal AI tasks (behavior classification accuracy, false alert rates, etc.)
- Collaborate with firmware and embedded engineers to ship models on resource-constrained hardware
- Monitor model performance in production, set up drift detection, and maintain model versioning and rollback processes
Qualifications & Skills
- Bachelor's or Master's degree in Computer Science, AI/ML, Electrical Engineering, or a related field
- 2+ years of hands-on experience training and deploying ML models in production
- Strong proficiency in Python; familiarity with C++ for performance-critical paths
- Deep experience with PyTorch and the Hugging Face ecosystem (Transformers, PEFT, TRL, Datasets)
- Practical experience with the modern post-training pipeline: SFT, preference optimization, and RL-based methods
- Understanding of modern model architectures: Mixture of Experts, long-context models, multimodal encoders
- Experience with inference optimization: quantization, speculative decoding, KV-cache management, batching strategies
- Familiarity with MLOps tooling: experiment tracking (W&B, MLflow), model registries, CI/CD for ML
- Knowledge of cloud GPU infrastructure (AWS, GCP) and cost-efficient training/serving strategies
- Strong problem-solving skills and ability to work in a fast-paced, cross-functional environment
- Background in animal behavior, veterinary science, or bioinformatics is a strong plus
Nice to Have
- Experience with video understanding models and temporal reasoning over long sequences
- Hands-on experience building agentic AI systems or tool-using LLM workflows
- Experience with structured outputs, function calling, and LLM-as-judge evaluation patterns
- Familiarity with on-device ML frameworks (Core ML, TensorFlow Lite, ONNX Runtime, ExecuTorch)
- Experience with synthetic data generation for training data augmentation
- Understanding of distributed training (FSDP, DeepSpeed) for large-scale model training
- Knowledge of AI safety, alignment techniques, and responsible AI practices
- Experience with Docker, Kubernetes, and production deployment pipelines
Why Join Us
- Work directly on AI that improves animal welfare and pet care. Your models have real-world impact on pets and their owners
- Full-stack AI role, from research and training to production serving and on-device deployment
- Access to GPU compute and cloud infrastructure for experimentation and training
- Collaborative team that values shipping over slides
- Competitive salary, equity, and benefits
We're building the future of intelligent pet care. If you want to push the boundaries of multimodal AI and apply it to a domain that genuinely matters, we want to talk.
