AI Architect - Remote

About the Role

This is an individual-contributor architecture role, deliberately hands-on. You will spend most of your time designing and building—prototyping, writing production code and infrastructure, and working directly in the codebases involved. You will also collaborate with product teams across the company, each with its own architects, roadmap, and deep domain expertise. Your influence will come from collaboration and demonstrated results rather than reporting lines.

Expect variety. A given week may include architecture and design work, implementation, embedding with a product team to resolve a dependency, and presenting ideas to internal stakeholders and customer design partners.

Responsibilities

• Design and build agent systems, including core reasoning loops, planning, state and memory management, tool interfaces, sub-agent delegation, and orchestration for multi-turn and autonomous agents.
• Make agent output trustworthy enough to act on, focusing on:
  • Confidence scoring and routing—determining what can proceed automatically, what warrants human review, and what requires direct handling.
  • Human-in-the-loop design to maintain professional judgment and accountability.
  • Audit trails and explainability sufficient for evidentiary and public-records scrutiny.
  • Evaluation practices: golden datasets, regression gates on model and prompt changes, staged and canary rollouts, and monitoring for drift.
  • Adversarial and failure-mode design: threat modeling for prompt injection and jailbreaks, containment of incorrect actions, and deterministic controls around probabilistic components.
  • Agent-level observability from the outset, tracking model calls, tools, inputs, and decision rationale.
• Make the existing portfolio accessible to agents. Build MCP servers, APIs, wrappers, and adapters to expose product capabilities; contribute to data pipelines and ingestion layers; and instrument applications to understand current usage patterns.
• Partner across product teams, working alongside architects and engineering leaders to design solutions or deliver work directly, then transferring knowledge for ongoing maintenance.
• Engage with customers, collaborating with design partner agencies and end users to ground work in real operational practice.
• Model current AI-assisted engineering practices, including skills, commands, hooks, MCP servers, sub-agents, specification-driven workflows, and deterministic quality gates (e.g., test suites, static analysis).

Requirements

• 10+ years building and shipping production software, including 3+ years at architect or staff-engineer scope spanning multiple teams or products.
• Current, daily, hands-on experience with agentic development tooling (e.g., Claude Code or equivalent), including building skills, commands, hooks, MCP servers, and sub-agents, and running specification-driven workflows.
• Experience designing LLM-based systems for reliable output: grounding and retrieval, structured output, confidence handling, hallucination control, human-in-the-loop patterns, and evaluation.
• Polyglot and stack-agnostic with depth in at least one of C#/.NET, Python, or TypeScript, and a demonstrated ability to quickly become productive in unfamiliar technologies, including legacy systems.
• Strong cloud architecture experience on a major public cloud (AWS preferred; Azure or GCP with translation ability accepted).
• Experience making legacy systems programmable via APIs, wrappers, adapters, or protocol-level interfaces (e.g., MCP).
• Observability as a first-class concern: distributed tracing, structured logging, metrics, and agent-level equivalents.
• Experience with containers, infrastructure-as-code (Terraform or equivalent), and CI/CD pipeline design.
• Track record of leading through influence rather than authority, collaborating with senior technical peers outside your reporting line.
• Clear written communication for specifications, architecture decision records, and design documents intended for both human and agent readers.

Preferred Qualifications

• Experience delivering an agent to production that takes action on a user’s behalf, including confidence gating, review, and rollback.
• Experience in regulated or high-consequence domains: public safety, healthcare, financial services, criminal justice, or government.
• Model-level work: fine-tuning, LoRA, distillation, local or self-hosted models, and cost/latency optimization.
• Data platform work: lakehouse layering, ingestion pipelines, and entity resolution across disparate systems.
• Multimodal experience: audio/speech, video, or sensor fusion.
• Personal or open-source projects demonstrating sustained, self-directed work with agent tooling.

Benefits

• Competitive compensation and a benefits package supporting work-life balance.
• Tuition reimbursement.
• Parental leave.
• Paid volunteer hours.
• Unlimited PTO.
• Flexible work environment (remote or in-office).

Additional Information

This role requires obtaining CJIS (Criminal Justice Information Services) clearance, which involves a comprehensive background check. U.S. citizenship or permanent residency is generally required for CJIS clearance. CentralSquare will support the clearance process if you are selected for hire.