Principal Architect, AI & Developer Productivity

PRINCIPAL ARCHITECT, AI & DEVELOPER PRODUCTIVITY

Location: Austin, TX (Hybrid, 2-3 days/week onsite).

Candidates must be authorized to work in the United States for the duration of employment without requiring current or future employer sponsorship of any kind.

About Us

TogetherWork is a $250M recurring revenue SaaS business with over 33 software applications serving 12+ vertical markets. We exist to help communities, organizations, and businesses thrive by delivering software solutions tailored to their unique needs.

With our new headquarters on South Congress in Austin, TX, we are entering an exciting phase of growth. We work hard and play hard. We love our customers, we innovate with purpose, and we won’t stop until we’ve delighted every client across the diverse ecosystem we serve.

About This Role

TogetherWork is seeking a Principal Architect, AI & Developer Productivity, reporting to [Reporting Manager Title], to own how AI accelerates the software development lifecycle across the portfolio. This is a hands-on leadership role for someone who has shipped AI-augmented engineering tooling at scale and can prove measurable improvements in developer throughput, software quality, and cycle time.

You will define and operationalize the AI-assisted development stack across the organization: IDE assistants, code review automation, test generation, security scanning, documentation, and release automation. You will set the standards, build the platform, and drive adoption across product teams. You will measure outcomes against DORA metrics and retire tools that do not produce returns, regardless of how fashionable they are.

This is not a research, thought leadership, or governance-only role. We are looking for someone who has actually deployed AI tooling into production engineering organizations and can show the metrics that prove it worked.

What You’ll Do

• Define and operationalize the AI-assisted engineering platform across the portfolio, covering IDE assistants, agentic coding tools (Claude Code, Cursor, etc.), code review automation, test generation, security scanning, documentation, and release automation.
• Architect a model- and vendor-agnostic abstraction layer so the organization is not locked into a single tool, model, or provider as the landscape evolves monthly.
• Establish reference architectures and golden paths for AI-augmented workflows that teams can adopt without forcing a single stack across all products.
• Establish acceptable use, IP protection, secret scanning, and data exfiltration controls for AI in the SDLC.
• Implement open source license scanning to prevent contamination from AI-generated code that reproduces GPL, AGPL, or other restrictive license material.
• Define audit trail and traceability standards: which AI tool wrote what code, what tests were generated, what was reviewed, and what was approved.
• Partner with Security, Legal, Compliance, and Risk to embed SOC 2, PCI, PII, SOX, data residency, and other regulatory requirements into the platform design, and support audit and risk assessment readiness.
• Embed AI-driven capabilities into CI/CD: automated pull request review, test synthesis, flaky test triage, vulnerability remediation, intelligent rollout, and incident analysis.
• Establish quality gates for AI-generated code, including coverage, mutation testing, security scanning, and license compliance before merge.
• Lead enablement across product teams: onboarding paths, paved roads, internal developer portal capabilities, and training for AI-assisted workflows.
• Treat developer experience as a product with clear roadmaps, success metrics, user research, and feedback loops.
• Distinguish real productivity from the illusion of productivity; AI tools inflate volume metrics, and commits and lines of code are unreliable in AI-native workflows.
• Report tool cost against measured outcomes, and make kill, scale, or replace decisions on tools that do not return $2 to $3 of value for every $1 of cost.
• Maintain an evaluation harness so new tools can be benchmarked against incumbents on real internal work, not vendor demos.
• Evaluate and select tooling across the market (GitHub Copilot Enterprise, Cursor, Claude Code, and emerging entrants), and negotiate enterprise terms in partnership with procurement.
• Make defensible build vs. buy decisions on AI components, frameworks, and pipeline integrations based on cost, security posture, switching cost, and outcomes.
• Stay current on emerging tools and models, and recommend platform evolution quarterly rather than annually.
• Bring acquired engineering teams onto the standard AI-augmented SDLC platform with a clear runbook, and provide structured recommendations on what to integrate, rationalize, or retire.
• Own the total cost of AI in the SDLC (license consumption, token spend, infrastructure, and developer time), including chargeback, cost ceilings, observability, and alerting.
• Build cost models for new tool rollouts that include training, change management, and ongoing platform support, not just license fees.
• Partner with engineering leaders, product, security, legal, and procurement to align platform direction with business strategy.
• Mentor senior engineers and engineering managers on AI-assisted development patterns and the discipline required to use them effectively.
• Communicate architecture decisions, trade-offs, and platform outcomes clearly to executive stakeholders, including the CTO and CFO.

What You’ll Bring

• 10+ years of proven expertise defining end-to-end solution architecture, including integration patterns, enterprise system architecture, developer experience, or engineering effectiveness roles, as a hands-on architect or senior engineer with direct shipping responsibility.
• Demonstrated production deployment of AI-assisted development tooling across multiple engineering teams, with measured outcomes. We will ask for specifics.
• Deep experience with modern CI/CD platforms such as GitHub Actions, GitLab CI, CircleCI, Jenkins, Buildkite, or equivalent.
• Hands-on experience with at least two of GitHub Copilot Enterprise, Cursor, Claude Code, Kiro, or comparable AI coding tools at organizational scale.
• Strong applied LLM knowledge: prompt design, context window management, RAG patterns, evaluation harnesses, model selection trade-offs, and cost and latency optimization.
• Experience designing controls for IP protection, open source license scanning, secret prevention, and data exfiltration in AI-assisted workflows.
• Strong systems engineering background: APIs, distributed systems, observability, and cloud-native architecture; AWS preferred.
• Experience operating at portfolio scale across 10 or more engineering teams with multiple technology stacks.
• Proven track record influencing without authority across protective engineering cultures and driving alignment across heterogeneous teams.
• Excellent written and verbal communication skills, including production of decision-quality technical documentation.
• Builder, not philosopher: has shipped, measured, iterated, and can show the receipts.
• Outcome obsessed: will retire a tool that does not produce returns even if it is fashionable, and will defend a boring tool that works.
• Pragmatic about risk: understands that IP protection, security, license compliance, and audit readiness are not afterthoughts in AI-augmented engineering.
• Comfortable in ambiguity: makes easily reversible decisions, avoids lock-in, and re-evaluates as needed.
• Product-minded: treats the internal developer platform as a product with users, roadmaps, and adoption metrics.
• Strategic and tactical: comfortable moving from executive briefing to code review in the same day.

Additional Requirements

• Ability to work onsite in Austin, TX 2-3 days per week as part of the hybrid schedule.

Why This Role

Impact: Define how AI accelerates software delivery across TogetherWork’s entire product portfolio, with outcomes measured against DORA metrics and ROI.

Challenge: Build a vendor-agnostic AI-augmented SDLC platform while balancing speed, security, IP protection, and compliance in a landscape that changes monthly.

Ownership: Set standards, select tools, own the total cost of AI in the SDLC, and make kill, scale, or replace decisions.

Collaboration: Partner with engineering leaders, Security, Legal, Procurement, and executives, including the CTO and CFO, to align platform direction with business strategy.

The Company offers a comprehensive employee benefits program, including:

• Comprehensive medical, dental, and vision plan options.
• Garner Health healthcare navigation and eligible expense reimbursement included with medical coverage.
• 100% employer-paid short-term and long-term disability coverage, Basic Life and AD&D insurance, and Teladoc access.
• Voluntary Accident, Critical Illness, Hospital Indemnity, Life, and AD&D insurance.
• Health Care and Dependent Care Flexible Spending Accounts (FSAs).
• 401(k) retirement plan with a 100% company match up to 4%.
• Flexible paid personal and vacation time; 10 paid sick days annually; and 10 company-paid holidays.
• Paid maternity leave, up to 12 weeks of paid parental leave, and up to 6 weeks of paid caregiver leave.
• Medical travel and fertility benefits, Employee Assistance Program, wellness benefits, employee engagement platform, and pet insurance.

Inclusion and Diversity

TogetherWork is an Equal Employment Opportunity Employer. We value diverse backgrounds, experiences, and viewpoints and do not make hiring or employment decisions on any basis protected by applicable local, state, or federal law.

Interview Process & Expectations

Interviews are designed to be interactive and conversational. Candidates are expected to respond based on their own experience and thinking during live interviews. To support a fair and consistent process, real-time AI tools or other external assistance may not be used to generate or guide interview responses.

Notes: I reused the About Us, benefits, EEO, and interview sections from your Senior Software Engineer template, since they’re more current than the PDF’s. The reporting line is a placeholder because the PDF doesn’t name one. I also dropped the CCPA links.