Pear VC
About the role
You’ll develop and deploy the ML systems that make Known work – from compatibility scoring and recommendations to natural language/voice interaction. You’ll work closely with platform engineers to shape the data foundation and with product engineers to deliver delightful user experiences.
Responsibilities
Design and implement profile matching algorithms (recommendations, search, personalization).
Build ML-specific pipelines for model training, evaluation, and inference at scale.
Research, experiment, and productionize agentic workflows to power voice agent and other similar functionalities
Deploy models into production and ensure monitoring, retraining, and evaluation loops.
Partner with the product team to run offline & online experiments to improve outcomes.
Requirements
4 years in applied ML engineering.
Expertise in Python with ML frameworks (PyTorch, TensorFlow).
Experience with recommendation/search/matching systems and LLM/agentic workflows.
Expertise in reinforcement learning is a big plus.
Familiar with model deployment, A/B testing, and monitoring in real-world systems.
Example Projects
Develop an embeddings-based retrieval re-ranking system for user compatibility.
Deploy a lightweight LLM agent for intake calls/voice matching.
Build an evaluation harness for offline model testing and online experimentation.
#J-18808-Ljbffr
Responsibilities
Design and implement profile matching algorithms (recommendations, search, personalization).
Build ML-specific pipelines for model training, evaluation, and inference at scale.
Research, experiment, and productionize agentic workflows to power voice agent and other similar functionalities
Deploy models into production and ensure monitoring, retraining, and evaluation loops.
Partner with the product team to run offline & online experiments to improve outcomes.
Requirements
4 years in applied ML engineering.
Expertise in Python with ML frameworks (PyTorch, TensorFlow).
Experience with recommendation/search/matching systems and LLM/agentic workflows.
Expertise in reinforcement learning is a big plus.
Familiar with model deployment, A/B testing, and monitoring in real-world systems.
Example Projects
Develop an embeddings-based retrieval re-ranking system for user compatibility.
Deploy a lightweight LLM agent for intake calls/voice matching.
Build an evaluation harness for offline model testing and online experimentation.
#J-18808-Ljbffr