Principal Software Engineer, AI
Principal Engineer · Principal · Full Time
Opens PermitFlow's application page
Role
What you'll do.
Principal Software Engineer, AI at PermitFlow is a high-impact staff/principal-level role leading the design and development of agentic AI infrastructure for a construction permitting platform serving the nation's builders. You will own the core AI assistant architecture, context layer, evaluation frameworks, and SDK standards while working directly with the VP of Engineering at a Series B company that has already delivered AI solutions across $20B in development projects. The role requires 10+ years of software engineering expertise with demonstrated experience shipping LLM-powered agentic products at scale, strong production systems thinking, and the ability to balance rapid iteration with rigorous evaluation and user-centered design.
Responsibilities
- Design and Own Agent Architecture: Architect the reasoning loop, planning and decomposition strategy, state and session model, and recovery behavior for long-running agentic tasks. Make critical decisions on how the AI assistant navigates complex construction workflows, handles task interruptions, and maintains consistency across multi-step permit submission processes.
- Build the Context and Retrieval Layer: Engineer retrieval augmented generation (RAG), memory management, and context-window optimization systems that enable the assistant to effectively work with construction documents, jurisdictional requirements, prior submissions, and customer-specific conventions. Design systems that maximize relevant information availability while minimizing token waste and latency.
- Design the Action Surface and Tool Framework: Define the tool interface layer that the assistant can call, implement granular permission models around actions, and establish human-in-the-loop checkpoints for expensive or irreversible operations. Create abstractions that balance autonomy with safety and auditability in a highly regulated domain.
- Establish Evaluation-Driven Iteration: Build comprehensive offline and online evaluation harnesses that measure task completion rates, identify failure modes, and track quality regressions. Create feedback loops that enable the entire engineering team to make data-driven decisions on agent improvements and prioritize impactful optimizations.
- Set Organization-Wide AI Engineering Standards: Establish foundational abstractions, SDKs, and code review standards that enable other engineers to build agentic features consistently. Serve as the technical authority on agentic patterns, best practices, and architectural decisions across the product organization.
- Drive User-Centric Product Development: Engage directly with permit coordinators and builders to understand real-world usage patterns, observe failure scenarios in production, and translate user insights into system improvements. Champion the voice of the end user in technical architecture decisions.
- Lead Technical Mentorship and Code Quality: Write production-grade code that establishes quality benchmarks for the team to follow. Mentor senior engineers on agentic infrastructure patterns and help scale the team's capability in building reliable AI systems.
Qualifications
What we look for.
Technical
Large Language Model Infrastructure
Deep expertise in building production systems with LLMs, including model selection, prompt engineering, fine-tuning strategies, and handling model outputs reliably in agentic contexts. Experience managing costs, latency, and quality tradeoffs in LLM-powered applications.
Agentic Architecture and Orchestration
Hands-on experience designing and implementing agent reasoning loops, planning algorithms, tool use patterns, and state management. Familiarity with orchestration frameworks like LangChain, CrewAI, or custom implementations. Understanding of ReAct, chain-of-thought, and other reasoning approaches.
Retrieval Augmented Generation (RAG)
Demonstrated ability to design and optimize RAG systems including vector databases, embedding strategies, retrieval ranking, context truncation, and integration with language models. Experience evaluating and improving RAG quality and relevance.
Evaluation Frameworks and Metrics
Experience building evaluation harnesses for AI systems including defining meaningful metrics, automating testing pipelines, and setting up feedback loops. Understanding of eval-driven development, red teaming, and identifying AI system failure modes.
Production System Design
Strong software engineering fundamentals including distributed systems thinking, API design, database architecture, monitoring, observability, and graceful degradation. Proven ability to move from prototype to production-grade systems that scale reliably.
Document Processing and Understanding
Experience working with document understanding, PDF extraction, form filling, and structured data extraction from unstructured content. Knowledge of document processing pipelines relevant to complex domain-specific workflows.
Education
Computer Science or Related Field
Bachelor's degree in Computer Science, Software Engineering, or equivalent field preferred. Advanced degree (Master's/PhD) in ML, AI, or related discipline a plus but not required if compensated by exceptional industry experience.
Experience
10+ Years Software Engineering
Minimum 10 years of professional software engineering experience with a track record of shipping production systems. Demonstrated progression in complexity and impact of projects owned.
AI Product Delivery at Scale
Recent, substantial experience (3+ years) shipping and iterating on LLM-powered or agentic products used by real customers. Proven ability to go beyond prototypes and deliver robust, user-facing AI systems.
Assistant or Agent Product Ownership
Hands-on experience building assistant or agentic products at meaningful scale, with direct responsibility for the core logic, quality, and reliability of the system. Exposure to real-world complexities, edge cases, and production issues in agentic systems.
Evaluation System Design
Track record of designing and implementing evaluation harnesses that materially changed team priorities and informed product decisions. Demonstrated ability to articulate what makes evaluations trustworthy and actionable.
Cross-Functional Leadership
Experience influencing product direction and technical decisions through demonstrated judgment, working code, and clear communication. History of earning authority through impact rather than title.
Skills
Required
LLM Product Development
Proven experience building customer-facing products powered by LLMs with focus on reliability, cost management, and user experience.
Agentic System Design
Deep expertise in designing agent architectures, tool use, planning strategies, and orchestration patterns for multi-step complex workflows.
Evaluation and Observability
Strong ability to design metrics, build evaluation pipelines, instrument systems comprehensively, and use data to drive prioritization.
Production-Grade Code
Ability to write clean, maintainable, well-tested code that scales. Strong software engineering discipline with attention to reliability and user impact.
System Architecture
Capability to design systems end-to-end including API contracts, data models, persistence layers, and operational concerns.
Technical Communication
Excellent written and verbal communication skills to articulate complex technical decisions, document architecture, and influence through clarity.
Preferred
RAG and Vector Database Experience
Nice to haveHands-on work with retrieval-augmented generation systems, vector embeddings, and semantic search in production environments.
Domain-Specific AI Applications
Nice to haveExperience building AI systems for highly specialized domains like legal, financial, regulatory, or construction tech where accuracy and liability are critical.
Multi-Agent Systems
Nice to haveFamiliarity with designing systems where multiple agents coordinate, decompose tasks, or operate in a specialized domain.
Construction or Regulatory Domain Knowledge
Nice to haveBackground in construction, permitting, compliance, or regulatory technology that brings domain context to technical decisions.
ML Infrastructure and Ops
Nice to haveExperience with ML infrastructure, prompt management systems, model versioning, experiment tracking, or ML ops tooling.
Founding or Early-Stage Experience
Nice to haveTrack record of building systems from scratch at startups or early-stage companies, comfortable with ambiguity and rapid iteration.
Tech stack
Languages
Frameworks
Databases
Tools
Other
Compensation
Pay and benefits.
Base·USD 250,000 – 300,000
Equity·Stock options
Benefits
Comprehensive Health Coverage
100% company-paid base medical, dental, and vision insurance for employees, plus health savings account (FSA) to reduce out-of-pocket costs.
Equity Compensation
Meaningful equity stake in a high-growth Series B company, allowing you to participate directly in company value creation alongside competitive salary.
Flexible Time Off
Unlimited PTO policy enabling you to balance work and life without tracking accrual, plus paid family leave for major life events.
401(k) Retirement Plan
Tax-advantaged retirement savings plan to help build long-term financial security.
Remote Work Support
Home office equipment stipend to set up a productive workspace, plus commuter benefits including pre-tax transit passes and parking support for NYC-based office days.
In-Office Amenities
Daily provided lunch and dinner at the NYC headquarters, removing friction from the workday and fostering team connection during hybrid work weeks.
Mission-Driven Work
Opportunity to build infrastructure that accelerates housing development, clean energy deployment, and infrastructure projects across America at a transformational scale.
Process
Interview steps.
- 01
Initial Conversation
Preliminary screening call with recruiter to discuss career background, interest in the role, and high-level qualification overview. Typical duration 20-30 minutes.
- 02
Technical Leadership Discussion
Detailed conversation with VP of Engineering or principal-level engineer covering your AI/agentic system architecture experience, past shipping decisions, and approach to technical leadership. Expected to discuss concrete examples of systems you've built.
- 03
System Design Deep Dive
Technical interview focused on designing agentic systems, evaluating architectural tradeoffs, and discussing how you'd approach core PermitFlow problems like agent reasoning, context management, and evaluation frameworks.
- 04
Product and User Understanding
Conversation exploring your approach to staying connected with users, gathering feedback on AI systems, and balancing technical purity with real-world constraints and customer needs.
- 05
Executive Alignment
Final conversation with founder or executive leadership to ensure cultural fit, vision alignment, and mutual interest in joining the PermitFlow mission of transforming construction acceleration.
Full posting
Original listing.
PermitFlow is redefining how America builds. We’re an applied AI company serving the nation’s builders, tackling one of the largest information challenges in the economy: understanding what can be built, where, and how. Our AI agent workforce helps the fastest-growing construction companies navigate everything from permitting and licensing to inspections and project closeouts – accelerating housing, clean-energy, and infrastructure development across the country.
Despite being a $1.6T industry, construction still suffers from massive delays, wasted capital, and lost opportunity. PermitFlow has already delivered unprecedented speed, accuracy, and visibility to over $20B in development, helping contractors reduce compliance time, de-risk projects, and scale with confidence.
As the U.S. enters a new capex supercycle across data centers, factories, housing, and renewables, joining PermitFlow means building the AI infrastructure at the core of every construction project driving the next wave of reindustrialization.
We’ve raised over $90M, most recently completing our Series B, from top-tier investors including Accel, Kleiner Perkins, Initialized, Y Combinator, Felicis, and Altos Ventures, with backing from leaders at OpenAI, Google, Procore, ServiceTitan, Zillow, PlanGrid, and Uber.
Our HQ is in New York City with a hybrid schedule (3 in-office days per week). We prefer NYC-based candidates or those open to relocation.
Role Overview
We're hiring a Staff/Principal Engineer to build PermitFlow's core product AI Assistant, infrastructure, and workflow agents. This is a role that will require UX and technical acumen to deliver high quality agents in a genuinely complex domain, while also up-leveling our core agentic infrastructure and data flywheel.
Submitting a permit means comprehending construction documents, extracting project details, confirming every requirement has been met and often filling in pdf forms. Errors compound across steps and a failed submission adds real cost for our customers as projects slow.
The ideal candidate will mix an exceptional history of product delivery with true experience building AI and Agentic infrastructure from the ground up. You ship quickly, instrument everything, hold a high quality bar, and stay close to the users of what you build.
This role reports directly to our VP of Engineering. We're hiring at Staff or Principal, with the level based on past experience and interview performance.
What You'll Do
Own the agent architecture. Design the reasoning loop, planning and decomposition strategy, state and session model, and recovery behavior for long-running tasks.
Build the context layer. Retrieval, memory, and context-window management over project documents, jurisdictional requirements, prior submissions, and customer-specific conventions.
Design the action surface. Define the tools the assistant can call, the permissions model around them, and the human-in-the-loop checkpoints for actions that are expensive or irreversible.
Make evaluation the engine of progress. Stand up offline and online eval harnesses that measure task completion and failure modes. Build the iteration loop that lets the whole team improve agent quality on evidence.
Set the bar for the org. Establish the abstractions, SDKs, and review standards other engineers build agentic features on.
Stay close to the customer. Sit with permit coordinators and builders, watch the assistant fail in real conditions, and bring that back into the system design.
What We're Looking For
You've shipped an assistant or agentic product at meaningful scale
Deep, current knowledge of the agentic stack (tool calling and tool design, context engineering, memory, retrieval, orchestration patterns, etc)
You’ve built eval harnesses that changed what a team worked on, and you can articulate what makes an eval trustworthy.
You are a genuinely strong software engineer, not an AI specialist who needs someone else to make it production-ready. You’ve owned services in production and you write code that other people are glad to inherit.
You operate with earned authority. You influence through demonstrated judgment, working code, and clear writing rather than title or mandate.
10+ years of software engineering experience with recent time spent building LLM-powered products.
What We Offer
Competitive salary and meaningful equity in a high-growth company
100% company-paid base medical, dental, and vision coverage for employees + healthcare FSA
401(k) savings plan
Unlimited PTO and paid family leave
Home office & equipment stipend
Daily in-office lunch and dinner provided
Commuter benefits (pre-tax transit and parking)
PermitFlow provides equal employment opportunities (EEO) to all employees and applicants for employment without regard to race, color, religion, sex, national origin, age, disability, genetics, sexual orientation, gender identity, gender expression, or family status, as protected by applicable law.
We are committed to a diverse and inclusive workforce and welcome people from all backgrounds, experiences, perspectives, and abilities. All employment decisions are based on merit, qualifications, and business needs.
Redirects to PermitFlow's application page.
Other roles