Staff Software Engineer | Semantic Layer, MetricFlow

This role is a core contributor position on MetricFlow, the open source semantic layer engine that turns metric definitions into SQL. The work named is designing features, optimising query generation, hardening the open source codebase and working with the external community of contributors and adopters. That last part matters: a meaningful share of the job happens in public, which is rare and shows up on a CV in a way closed source work cannot. Technically it sits in compiler territory, since the requirements ask for experience designing systems involving parsing, compilation or transformation, naming query planners, DSL compilers and type systems. The posting publishes no years figure, so we recorded it blank. Note the title says Staff while the body says Senior.

Fivetran publishes no experience figure anywhere in this posting, so we have left the field blank rather than guess, and this listing sits in no experience band on the page. Note also an internal inconsistency worth knowing about: the board title says Staff Software Engineer while the description text says the company is hiring a Senior Software Engineer. Clarify the level and band early, since that gap can be worth a lot in compensation. Published requirements: strong professional software engineering experience with a track record of shipping production systems; deep proficiency in Python including modern typing, testing practice and packaging; strong SQL fundamentals and a working understanding of how analytical query engines and data warehouses execute queries; and experience designing systems that involve parsing, compilation or transformation, with query planners, DSL compilers and type systems given as examples. The role is on the Semantic Layer team in Bangalore, as a core contributor to MetricFlow. Described as a high ownership role where you write production code, drive technical decisions and help shape long term direction, with responsibilities spanning feature design, query generation optimisation, hardening the open source codebase and partnering with the community of contributors and adopters around it. Why this is technically interesting: a semantic layer takes metric definitions expressed once and compiles them into correct SQL across different warehouses and query shapes, which is genuinely a compiler problem with all the usual concerns, an intermediate representation, correctness under composition, and generated output that must be both right and efficient. The open source element means your commits and design discussions are public. Same employer caution as our other Fivetran listing: this company flags some postings with an internal hiring priority status meaning they are not open externally even though the page loads normally. We checked this requisition and it carries no such flag.