Senior Software Engineer
Job description
Senior Software Engineer at Mercury
About the role
Join Mercury to build the future of financial technology. You will focus on evolving Command, our LLM-powered financial assistant, making it more capable and intuitive for our customers. This role involves shaping how users interact with their finances through plain language.
Key facts
What you'll do
Develop and launch new features for Command, including defining and implementing instruction sets that enable the AI to perform financial tasks.
Architect and build multi-step agentic workflows, designing the underlying systems for complex, sequential AI interactions.
Collaborate with backend engineers to establish data contracts for new functionalities, bridging business logic and AI models.
Take ownership of new capabilities from initial prompt design through to the user interface components.
Manage and enhance the core prompt architecture of Command, including system prompts, skill management, and compliance layers.
Optimize model behavior for speed, cost, and effectiveness, adjusting parameters like reasoning effort and caching.
Stay informed about advancements in AI models and integrate new knowledge into Command's development.
Create and expand automated testing frameworks to ensure the quality and reliability of new features.
Work with product and compliance teams to define success criteria and build tests that validate functionality.
Monitor and ensure the ongoing performance and quality of shipped features.
Requirements
Possess at least 7 years of software engineering experience, with substantial practical knowledge in building and scaling LLM-powered applications in a production environment.
Demonstrated experience in scaling an LLM product beyond its initial release, addressing production reliability and performance challenges.
Proven ability in designing agentic systems and architecting multi-step workflows that are dependable, understandable, and safe for user operations.
Experience building evaluation infrastructure and creating meaningful tests that assess product functionality beyond plausible model output.
A clear understanding of the practical considerations in LLM deployment, including latency, cost, compliance, and production failure modes.
A thoughtful perspective on building trustworthy AI products, not just impressive ones.
Proficiency in TypeScript and a willingness to learn Haskell for backend tool development, or existing expertise in both.
Ability to work across the entire stack of an AI product, from system prompts to frontend streaming interfaces.
A history of mentoring other engineers and improving the technical capabilities of a team.
Nice to have
Experience with building AI products that are trustworthy and reliable.
Skills & tools
TypeScript
Haskell
Practical notes
US employees: Base salary range of $200,700 - $250,900.
Canadian employees: Base salary range of CAD 189,700 - 237,100.
Compensation includes base salary, equity, and benefits. Ranges are competitive and updated based on industry data. Offers are determined by experience, expertise, location, and internal equity.
Banking services are provided by Choice Financial Group and Column N.A., Members FDIC.