AI Architect – Mathematical Reasoning Systems

About KAITLab KAITLab is building a next-generation learning intelligence platform designed to understand how students think in mathematics—not merely whether answers are correct. Our system analyzes step-level mathematical reasoning from students in grades 3–10 to model conceptual understanding, procedural fluency, and mastery progression aligned with the Common Core State Standards. The objective is to develop a reasoning-aware AI system capable of interpreting student thinking and guiding adaptive instruction in real time. We are seeking a Founding AI Architect to design the core reasoning and personalization architecture that powers this intelligence layer. This is not a prompt-engineering role. The work centers on designing algorithm-driven machine learning and reasoning systems capable of interpreting structured mathematical work, identifying misconceptions, modeling evolving knowledge states, and determining optimal instructional responses. The architecture developed in this role will serve as the decision engine behind a new generation of adaptive mathematics learning technology. Research Orientation This role sits at the intersection of AI system architecture, mathematical reasoning, and cognitive modeling. We are particularly interested in candidates motivated by challenging problems involving structured reasoning, algorithmic learning, and interpretable AI systems. The successful candidate will help define foundational architecture for a platform exploring how AI can model student reasoning, misconception formation, and conceptual learning in mathematics at scale. This role is intended for individuals with deep expertise in algorithmic AI system design and structured reasoning systems who are interested in shaping foundational architecture for an emerging AI-driven learning platform. The Core Technical Challenge Design an AI architecture capable of: · Interpreting step-by-step student mathematical reasoning · Detecting procedural and conceptual errors · Classifying mathematical misconceptions · Modeling evolving student knowledge states · Determining mastery thresholds · Driving adaptive instructional decisions The resulting system will likely involve hybrid algorithmic architectures combining classical algorithms with modern machine learning approaches, including: · Symbolic reasoning · Probabilistic modeling · Sequential learning models · Machine learning classification · Structured mathematical parsing Responsibilities As Founding AI Architect, you will lead the design of the KAITLab reasoning intelligence engine, including: · Step-level reasoning evaluation models · Misconception detection and classification frameworks · Student knowledge-state modeling systems · Mastery progression modeling · Confidence estimation and intervention thresholds · Hybrid symbolic–neural architectures for mathematical reasoning Additional responsibilities include: · Translating research-grade models into implementable engineering specifications · Defining evaluation metrics for: o Step-level reasoning accuracy o Misconception detection recall o Mastery prediction reliability o Adaptive intervention effectiveness · Producing technical architecture documentation and system roadmaps · Working closely with the founding team to guide development of the next generation KAITLab AI engine Required Qualifications · PhD or equivalent research experience in one of the following: o Artificial Intelligence o Machine Learning o Applied Mathematics o Computational Cognitive Science o Educational Data Mining · Demonstrated experience designing reasoning-aware AI systems or algorithm-driven ML architectures · Strong background in probabilistic modeling · Experience with sequential or latent-state models, such as: o Bayesian Knowledge Tracing o Hidden Markov Models o Reinforcement Learning o Graphical Models · Experience with structured input representations, symbolic reasoning systems, or neuro-symbolic architectures · Proven ability to translate research concepts into production-grade system architecture Preferred Background Experience in one or more of the following areas is highly valuable: · Intelligent tutoring systems · Educational AI · Cognitive modeling · Mathematical reasoning systems · Knowledge tracing · Structured symbolic computation Experience working in early-stage research environments or venture-backed startups is helpful but not required. Expected Outcomes (First 6–9 Months) The initial engagement is expected to deliver: · Architecture audit of the existing system · Design of Version 2 reasoning intelligence architecture · Formal misconception taxonomy framework · Step-level reasoning evaluation models · Student knowledge-state modeling system · Adaptive instruction decision framework · Engineering roadmap for implementation Engagement Structure and Collaboration This is a fractional role (approximately 10–20 hours per week) designed for a senior researcher or architect who can guide core system design while collaborating closely with the internal team. Because the AI architecture will directly influence the product roadmap, regular collaboration with the team is required. Candidates should expect: · A small number of scheduled collaboration hours during core team availability each week · Typically 3–5 hours of synchronous collaboration during overlapping work hours · Core collaboration windows generally fall between 10:00 a.m. and 4:00 p.m. Eastern Time KAITLab operates across multiple time zones, and meeting schedules are coordinated to accommodate distributed team members when possible. Final expectations regarding working hours and collaboration cadence will be mutually agreed upon during the hiring process. How to Apply Please submit: · CV or resume · A brief description of reasoning-based AI systems you have designed · Links to publications, technical reports, or GitHub repositories (if available) Applications will be reviewed on a rolling basis. Applications must be submitted via email to terri.dennis@kaitlab.com.Indeed Apply submissions will not be reviewed. Pay: $50.00 - $60.00 per hour Expected hours: 10.0 – 20.0 per week Work Location: Remote