About The Role
Build the next-generation DAX-like formula engine that turns complex evaluation contexts into efficient SQL on modern data warehouses. As a Staff Engineer on the BI Engine team, you will own the core compiler and runtime for our standalone dax_engine — spanning the parser, metadata layer, query planner, SQL translation backends, and client connectivity — and deliver a BI experience that faithfully mirrors native models while integrating tightly with Arrow- and DuckDB-based storage.
Responsibilities
Engine Ownership:
- Own the Spotonix BI engine end to end BI catalog and metadata, formula-language grammar and parser, query planner and optimizer, SQL generation,execution layers, client connectivity,session management, and multi-dialect SQL support.
- Drive observability for plan generation, SQL issuance, and cache effectiveness, partnering with integrations teams to support additional warehouses, metadata loaders, and client APIs.
Compiler & Query Planning
- Design and optimize the query compiler that lowers expressions into an internal multidimensional algebra, applies rule- and cost-based transformations, orchestrates algebraic plans, manages evaluation contexts, and enforces lineage propagation before emitting SQL.
- Evolve the execution runtime translate complex algebra into SQL with maximum warehouse pushdown and expose the instrumentation hooks needed by clients and regression suites.
BI Algebra Design
- Own the BI algebra itself define and implement the suite of DAX,MDX-style operators (hierarchy navigation, cube operators, table manipulations, iterators) and ensure each algebraic primitive ships with tests and SQL strategies.
- Expand DAX function coverage and time-intelligence patterns while maintaining correctness and semantic fidelity across all supported warehouse backends.
Technical Leadership & Collaboration
- Mentor engineers on query-compilation fundamentals, context transition debugging, and deterministic developer workflows (diagram generation, regression fixtures, CLI ergonomics).
- Collaborate with UI and data-modeling teams, writing dev notes and codifying manual plans into robust, repeatable test fixtures.
Qualifications
Required
- 10–15 years building database engines, query compilers, or analytical runtimes with deep experience in SQL planning and execution internals (costing, rewrites, vectorized operators, columnar caches).
- Expert-level Python and ease working in large metaprogrammed codebases; C++, Rust, or Scala for runtime paths is a plus.
- Demonstrated ownership of multidimensional or BI algebra design — defining operators, invariants, and execution strategies across catalogs, hierarchies, and cube semantics.
- Hands-on knowledge of DAX-like BI languages (MDX, LookML metrics, dbt semantic models, etc.) and their filter,context semantics, especially CALCULATE, context transitions, and time-intelligence functions.
- Proven track record translating high-level DSLs into relational algebra or SQL, including dependency tracking, lineage-aware filter propagation, and multi-stage compilation pipelines.
- Familiarity with Arrow, DuckDB, or similar columnar storage engines, plus the ability to reason about concurrency, caching, and resource management in a long-running service.
Preferred
- Experience building client,session infrastructure (CLI, APIs) and managing multi-dialect SQL connectors.
- Comfortable operating in product-oriented environments writing dev notes, codifying manual plans into tests, and collaborating with UI and data-modeling teams.
Nice to Have
- Experience ingesting PBIX or Tabular models, XML,A-style protocols, or enterprise BI ecosystem tooling.
- Background in time-intelligence modeling, date relationship functions, or working-days calendars.
- Contributions to open-source database engines, query planners, or columnar execution projects.
- Passion for developer tooling GraphViz visualizations, SQL trace explorers, CLI experiences, and regression harnesses for semantic languages.
What Success Looks Like
You deliver a DAX-like engine that faithfully reproduces complex evaluation semantics — context transitions, lineage-aware filters, manual rowset DAGs — while compiling to warehouse-native SQL that hits performance SLAs.
Query authors and integration teams gain a predictable evaluate() surface, cache hit rates climb, SQL traces look hand-written, and the coverage of DAX functions plus time-intelligence patterns keeps expanding without sacrificing correctness.
Spotonix is an equal opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.
