MLE, ML Platform

Job DescriptionJob DescriptionAbout zaimler AI agents can't reason over data they don't understand. Enterprise data today is fragmented across dozens of systems with no shared context, meaning, or structure, and that's why most enterprise AI is failing. The shift from copilots to autonomous agents is creating an entirely new infrastructure layer, and we're building it. zaimler is the context infrastructure for the agentic era: a platform that automatically discovers domain knowledge, maps relationships, and gives AI agents the semantic understanding to operate with precision at scale.

Imagine knowledge graphs that support real-time inference, built for systems that need to reason, not just retrieve. zaimler was founded by Biswajit Das (ex-VP Engineering, Truera), a Data Infra veteran and former Chief Architect at Visa, and Sofus Macskassy (ex-Director of Engineering, LinkedIn), who built one of the largest knowledge graphs in production in the industry at LinkedIn. We're growing and deploying with major enterprises across insurance, travel, and technology. If you want to build infrastructure that the next decade of enterprise AI runs on, we'd love to talk.

About the Role You’ll join our ML team focused on turning raw enterprise data into structured, contextualized knowledge graphs and embeddings. You’ll develop novel and highly scalable algorithms for ML and data engineering to make our overall system more efficient, experiment with new approaches for distilling large models into smaller, more efficient ones; improve retrieval, ranking, and reasoning performance through feedback loops; and prototype methods that help LLMs extract and act on real-world knowledge. We're looking for someone who thrives on iteration, cares about building with rigor, and is hungry to learn from some of the best engineers and researchers in the field.What You’ll Be DoingBuild and maintain training infrastructure, feature stores, and model serving pipelinesOptimize LLM inference performance — compute efficiency, memory management, latency, and throughputRead, debug, and contribute to LLM runtime and supporting library code (Rust and/or C++)Deploy and manage models at scale using tools like vLLM and BasetenArchitect scalable pipelines for model training and serving across GPU infrastructureCollaborate with ML and data engineers to ensure the platform meets research and production needsPrior ExperiencePhD in CS, ML, or a related field or MS with 4+ years of relevant industry experienceBackground in LLM optimization: inference efficiency, quantization, memory layout, or serving performanceAbility to read, navigate, and debug LLM source code and underlying runtime librariesComfortable in Rust and/or C++ at the systems level; strong Python requiredStrong algorithmic fundamentals — data structures, complexity, distributed systemsHands-on experience with model serving infrastructure (vLLM, Baseten, Triton, or similar)Experience setting up and scaling ML pipelines end-to-endNice to HaveFamiliarity with feature store design and managementExperience with GPU cluster management and optimizationContributions to open-source ML infrastructure or LLM toolingExperience with Ray, ONNX, TensorRT, or similar optimization and serving frameworksUnderstanding of transformer internals and attention mechanisms at the implementation levelWhy JoinA rare chance to be a founding engineer shaping both company and product direction.Competitive salary, benefits, and meaningful equity.Work alongside engineers and researchers from LinkedIn, Visa, Meta, and Branch.Onsite culture in San Mateo, designed for deep collaboration and high-velocity building.Full benefits package (Medical, Dental, Vision, 401k).We sponsor H-1B visas and assist with immigration processes.We value builders over résumés.

If this role excites you but you don't check every box, we still want to hear from you. zaimler is an equal opportunity employer.We may use artificial intelligence (AI) tools to support parts of the hiring process, such as reviewing applications, analyzing resumes, or assessing responses. These tools assist our recruitment team but do not replace human judgment. Final hiring decisions are ultimately made by humans.

If you would like more information about how your data is processed, please contact us.