Granica is looking for a Software Engineer – Foundational Data Systems for AI to build the metadata substrate, adaptive engines, and intelligent data layouts that power efficient, structured AI at exabyte scale. You will translate research algorithms into production-grade implementations, working at the intersection of fundamental research and enterprise impact.

What You'll Do

Design and implement the global metadata substrate supporting time-travel, schema evolution, and atomic consistency across massive tabular datasets.
Build adaptive engines that reorganize data autonomously, learning from access patterns and workloads to maintain efficiency with minimal manual tuning.
Develop and refine bit-level encodings, compression, and layout strategies to extract maximum signal per byte read.
Contribute to distributed compute systems that scale predictively and adapt to dynamic load.
Translate new algorithms in compression and representation from research into production-grade implementations.
Design and optimize data paths to minimize time between question and insight, enabling faster learning for both models and humans.

What We're Looking For

Foundational understanding of distributed systems: partitioning, replication, and fault tolerance.
Experience or curiosity with columnar formats such as Parquet or ORC and low-level data encoding.
Familiarity with metadata-driven architectures or data query planning.
Exposure to or hands-on use of Spark, Flink, or similar distributed engines on cloud storage.
Proficiency in Java, Rust, Go, or C++ and commitment to clean, reliable code.
Curiosity about how compression, entropy, and representation shape system efficiency and learning.
A builder’s mindset—eager to learn, improve, and deliver features end-to-end with growing autonomy.

Nice to Have

Familiarity with Iceberg, Delta Lake, or Hudi.
Contributions to open-source projects or research in compression, indexing, or distributed systems.
Interest in how data representation influences AI training dynamics and reasoning efficiency.

Technical Stack

Parquet, ORC, Spark, Flink, Java, Rust, Go, C++, Iceberg, Delta Lake, Hudi

Team & Environment

You will partner closely with the Granica Research group led by Prof. Andrea Montanari (Stanford). The environment emphasizes deep technical work, minimal bureaucracy, and a shared mission.