← Back to Projects
Accelerating Graph Indexing for ANNS
Mentored by: Next Silicon
Optimized graph-based indexing for approximate nearest neighbor search
C++
Graph Algorithms
Vector Search
HNSW
GPU Acceleration
Benchmarking
Description
A high-performance indexing system for approximate nearest neighbor search (ANNS) using graph-based data structures. Accelerates similarity search in high-dimensional spaces with improved query performance and reduced memory footprint. Features include incremental indexing, dynamic updates, and support for various distance metrics.
Team Members
Cohort: Embedded Systems Bootcamp 2025 (Embedded)
Responsibilities:
Improved performance through OpenMP Multithreading and parallel computation using SIMD.
Optimized memory access patterns (cache locality, SoA layout, prefetching).
Integrated Machine Learning models (e.g., KMeans and PCA).
Worked in a Linux environment, using CMake for builds and running tests with Google Test.
Built a CI process with GitHub Actions and ran Google Benchmark for comparison against HNSW.
Performed bottleneck analysis and optimization using Intel VTune profiler.
...and more contributions not listed here
Responsibilities:
parallel execution using SIMD
Integrated Machine Learning models (e.g., KMeans and PCA).
Profiled system performance with Intel VTune and flame graphs to identify bottlenecks.
Implemented comprehensive unit tests with Google Test to ensure code reliability.
Built a CI process with GitHub Actions and ran Google Benchmark for comparison against HNSW.
Worked according to standard development methodologies, including professional code review.
...and more contributions not listed here