Senior Principal Engineer

Saige Partners•Folsom, CA

2d•Onsite

About The Position

This is a contract position for a Senior Principal Engineer located in Folsom, CA. The role focuses on enabling vector database frameworks on an FPGA-accelerated platform, modifying software stacks for offloading computations to FPGA-based storage or accelerator devices, and designing host-side software paths for computation offload. The engineer will develop and optimize I/O submission/completion paths using modern Linux interfaces, analyze VectorDB execution flows to identify offload opportunities, and debug full-stack issues. Saige Partners is a growing technology and talent company committed to helping professionals build careers.

Requirements

Bachelor’s or Master’s degree in Computer Science, Computer Engineering, Electrical Engineering, or related field
Strong programming skills in C/C++ and experience with performance-oriented system software development
Hands-on experience with vector databases, similarity search, or ANN frameworks such as DiskANN, Milvus, FAISS, or related systems.
Good understanding of Linux systems programming, storage I/O, and asynchronous I/O models
Familiarity with NVMe protocol, command structure, submission/completion queues, and storage performance concepts.
Experience analyzing and modifying open source software stacks for performance optimization and platform integration
Strong debugging and profiling skills across user-space and system-level software.

Responsibilities

Enable vector database frameworks such as DiskANN, Milvus, and related retrieval/search pipelines on an FPGA-accelerated platform
Modify and optimize software stacks so that distance computation and related vector operations can be offloaded to FPGA-based storage or accelerator device
Design and integrate host-side software paths to issue custom or vendor-specific NVMe commands for computation offload
Develop and optimize I/O submission/completion paths using modern Linux interfaces such as io_uring
Analyze existing VectorDB execution flows and identify where offload can be inserted with minimal software disruption
Debug full-stack issues spanning application software, libraries, kernel/user-space I/O paths, NVMe protocol behavior, and FPGA device responses.