Principal Architect – AI

This is a fully remote position, open to applicants in United States.

📋 Description

• Take charge of the design and development of Brightside’s AI agent platform.

• Create and oversee the multi-agent orchestration layer, which includes the session supervisor, domain clusters, escalation arbiter, and conversation manager.

• Establish agent routing logic, context sharing, tool utilization, and latency budgets across more than 18 agents.

• Construct the real-time pipeline linking Amazon Connect through Contact Lens, Kinesis, Lambda, Bedrock, and WebSocket push to the Financial Assistant panel.

• Manage dual-mode agent architecture for both human-assisted coaching and autonomous client interactions.

• Determine model selection and routing across low-latency and reasoning-intensive agent tiers.

• Control prompt architecture, encompassing structured agent prompt standards, runtime domain knowledge injection, and conditional output formatting.

• Develop and manage the evaluation harness for schema validation, behavioral regression testing, and latency profiling.

• Design the domain knowledge base across eight domains and 49 use case types.

• Oversee the AI platform data model for client financial profiles, cases, goals, options, outcomes, and financial impact measurement.

• Architect the CRM synchronization layer, which includes real-time field writes post-Financial Assistant confirmation and logging of human corrections.

• Define canonical data models across clients, employers, financial products, and outcomes.

• Maintain data integrity across Aurora PostgreSQL and Aurora MySQL.

• Manage AWS Bedrock-native inference infrastructure, including asynchronous invocations, Knowledge Bases, and RAG pipelines.

• Design the session state layer utilizing ElastiCache Redis and DynamoDB.

• Establish AI governance standards for model evaluation, monitoring, explainability, and compliance.

• Assess and guide decisions regarding model providers, orchestration frameworks, and platform tooling.

• Collaborate directly with Jacky Chiu and the engineering team to architect and deliver production systems.

• Troubleshoot Bedrock invocations, analyze latency spikes, and review schema modifications.

⛳️ Requirements

• At least 10 years of experience in software engineering or systems architecture, with considerable hands-on experience in the last three years.

• Experience in production environments building real-time AI systems that operate within a per-turn, sub-second latency budget.

• Proficient with Amazon Bedrock, AWS Lambda, and Kinesis.

• Experience in multi-agent system design, including orchestration patterns, agent handoffs, context sharing, and structured output validation.

• Strong foundation in data architecture, including relational modeling, event-driven systems, and CRM data patterns.

• Familiarity with regulated environments such as fintech, financial services, or healthcare, including knowledge of privacy and compliance requirements.

• Comfortable working in a small, fast-paced team environment where you will design and build.

• Experience in developing voice or conversation intelligence systems.

• Practical experience with Amazon Connect and Contact Lens.

• Background in conversational AI product companies such as Cresta, Observe.AI, Cogito, Replicant, or similar.

• Proficient in LLM prompt engineering at a systems level, including evaluation harnesses, versioned prompt governance, and regression testing.

• Previous experience as a lead architect at a Series B-D company.

• Capable of forming, articulating, and updating technical opinions.

• Focused on production outcomes.

• Comfortable troubleshooting Bedrock invocations, analyzing latency spikes, and reviewing schema changes.

• Able to balance experimentation with rigor in a regulated fintech environment.

🏝️ Benefits

• Inspiring career opportunities.

• Excellent work/life balance.

• Meaningful work that directly impacts working families.

• Opportunity to witness the growth of a start-up from an insider’s perspective.