• Design, build, and deliver large-scale and real-time systems with high quality, scalability, and operability in mind. • Collaborate with product management, data science, and other engineering teams to bring end-to-end solutions to our customers. • Review and provide feedback on technical designs and code changes to engineering peers. • Proactively identify areas of technical debt and propose solutions to address them. • Take operational ownership of the system by using monitoring and observability tools to identify and respond to issues. • Mentor and coach team members to raise the technical bar on the team.

• Develop and optimize RAG pipelines: Build advanced Retrieval-Augmented Generation (RAG) systems that query multiple data sources, such as PDFs and video transcripts, to ensure accurate responses with broad content coverage. • Automate quality evaluation: Implement "LLM-as-a-Judge" pipelines to audit the quality of AI responses at scale. • Build AI personalization: Develop and implement long-term conversational memory systems (using frameworks like LangChain) so the system can remember context and interaction histories. • Architect multi-agent systems: Design and build an architecture of specialist agents with an intelligent router that directs requests to the most appropriate agents. • Work with LLMs: Use and optimize the latest large language models (GPT, Claude, etc.), applying prompt engineering techniques to maximize response quality. • Contribute to knowledge vectorization: Support the strategic "Ubiquitous AI" initiative by helping to structure and vectorize Gran's entire academic repository to make it accessible to AI. • Define data contracts: Collaborate with software developers to define the APIs and data contracts required to serve AI models efficiently and with low latency.

• Build greenfield Python applications using agentic methods, including AI agent workflows with tools such as Claude Code, LangChain, or equivalent frameworks • Architect the scaffold layer that coordinates multi-agent pipelines, including role separation, context injection, and output validation • Integrate agentic systems with the GitHub-based knowledge store that serves as the org-wide context layer for engineering work • Evaluate and adopt new agentic tooling and LLM capabilities as they emerge, bringing structured recommendations to the team • Design and implement services, RESTful and event-driven APIs, and data access layers across Python backends and React frontends • Contribute to integration between internal services and external platforms, including third-party APIs relevant to IR workflows and business operations • Write code that other engineers can read, extend, and trust in production • Own and operate AWS infrastructure across all environments (dev, staging, production), including Kubernetes, Terraform, and CI/CD via GitHub Actions. Engineers here own what they ship, from code to infrastructure — that’s a feature, not a burden • Build monitoring, alerting, and observability so production issues surface early and resolve fast • Ensure infrastructure aligns with Surefire’s security and compliance requirements. We’re a cybersecurity company; security is a design constraint, not a review step • Identify weaknesses in the current architecture and propose pragmatic improvements with clear rationale.

• Lead high-impact, cross-functional projects across Treasury engineering, operations, compliance, and risk teams to build stable and scalable systems and processes • Play a significant role in defining the technical architecture and experience of the Treasury product • Collaborate with product, design, and operations to ship thoughtful features for our customers and internal users • Feel a strong sense of product ownership and actively seek responsibility – we want someone excited to help shape and build Mercury’s future • Join a team of experienced and junior product engineers managing billions of dollars of customer assets