Ajit Mistry
Software Engineer, Sys SW at NVIDIA
Zurich, Zurich, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolAjit Mistry is a software engineer specializing in systems and ML infrastructure with six years of hands-on experience accelerating vector search and inference workloads. Currently on NVIDIA’s Sys SW team after a DL Inference Safety internship there, he previously drove major performance gains at Weaviate—implementing SIMD and assembly optimizations that sped distance calculations and indexing by an order of magnitude and integrating GPU vector search (cuVS) to boost query throughput up to 70x. Comfortable across C++, Go, and low-level assembly, he bridges research papers (e.g., DiskANN) to production by reimplementing and shipping high-performance components. He combines strong open-source contributions—most notably SIMD kernels merged into the popular Weaviate vector database—with an ETH Zurich CS education. Colleagues rely on him to translate hardware-aware optimizations into measurable system improvements and productive partnerships with vendor teams. An appetite for full-stack demos and benchmarks means his work is both demonstrably fast and easy for others to adopt.
6 years of coding experience
1 year of employment as a software developer
Abitur, Abitur at Heinrich-von-Gagern-Gymnasium FrankfurtBachelor, Informatik, Bachelor, Informatik at ETH Zürich
German, English
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (6)
vector-search10
simd10
vector-database10
go10
c178
c118
Programming languages (9)
Github contributions (5)
weaviate/weaviate
Mar 2024 - Mar 2025
Weaviate is an open-source vector database that stores both objects and vectors, allowing for the combination of vector search with structured filtering with the fault tolerance and scalability of a cloud-native database.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:6 reviews, 16 PRs, 121 pushes in 1 year
Contributions summary:Ajit implemented SIMD (Single Instruction, Multiple Data) optimizations for the Hamming and L2 distance calculations within the Weaviate vector database, specifically targeting ARM64 and AMD64 architectures. They wrote assembly code using GOAT to generate optimized routines, enhancing performance. Additionally, the user worked on integrating and testing these optimized distance calculations within the Weaviate codebase.
approximate-nearest-neighbor-searchsemantic-search-enginehnswfaultsimilarity-search
ajit283/cuvs
Jun 2024 - Mar 2025
cuVS - a library for vector search and clustering on the GPU
Contributions:18 PRs, 147 pushes, 14 branches in 9 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.