Real-Time AI Database Engineer

abra R&D is looking for a Senior Embedded Engineer! abra R&D is looking for an Embedded Engineer to help build the first real-time database purpose-built for AI agents at scale. Designed around time-series and unstructured data, it leverages a custom storage format optimized for append-heavy, real-time workloads. Joining the team means working on a high-performance execution engine built with vectorized execution, SIMD, and cache-efficient memory layouts, enabling extreme low-latency and high-throughput performance at scale. The system is engineered to fully utilize modern hardware, with deep optimization across CPU cache layers (L1 / L2 / L3), pushing the limits of real-time data processing. What You’ll Do Design and implement core components of the database engine (query engine, execution engine, storage engine) Build vectorized execution pipelines optimized for SIMD Design and evolve a time-series optimized storage engine (custom on-disk + in-memory formats) Work on unstructured / event-driven data representations and efficient indexing/querying strategies Own memory layout, compression, and data defragmentation Develop cache-aware / cache-efficient data structures with deep understanding of CPU cache behavior Implement distributed data primitives: sharding, partitioning, replication, and data locality Profile and optimize performance at the CPU level (cache misses, branch prediction, memory bandwidth)

Requirements: 7+ years of experience in high-performance systems programming (C/C++) Deep understanding of computer architecture (CPU pipelines, cache hierarchies, memory access patterns) Strong experience with low-level optimization and profiling tools Proven knowledge of multithreaded development (lock and lockfree) Expertise in algorithms and data structures, especially cache-aware designs Experience in one or more of the following: Database internals (query engines, storage engines, query planners/optimizers) Time-series or real-time data systems High-performance systems (trading systems, game engines, networking stacks, compilers) Distributed systems (sharding, partitioning, consistency models) Strong Plus Experience with vectorized execution engines (e.g., DuckDB-style processing) Experience designing custom storage formats or low-level data layouts Experience handling unstructured or semi-structured data at scale Background in query optimization and execution planning