Agentic AI Engineer

Agentic AI Engineer

Hybrid/Remote

Role Purpose

TrekAI is reinventing education through AI. As a high-growth, mission-driven startup, we build technology that directly impacts teachers, students, and school systems.

As TrekAI’s Agentic AI Engineer, you will design, build, and ship the models, algorithms, and policies that power adaptive, personalized, and explainable learning experiences. Working alongside senior AI staff, product, and engineering, you will translate cutting-edge research in LLMs, causal inference, and reinforcement learning into scalable, production-grade AI systems that reach real classrooms.

This is a rare opportunity for an early-career engineer with a deep theoretical and mathematical foundation to help pioneer next-generation agentic learning systems and turn research into measurable educational outcomes.

Key Responsibilities

• Build and evaluate multi-agent / agentic architectures for reasoning, retrieval-augmented generation (RAG), and tutoring workflows.
• Implement and fine-tune LLM-based systems — prompting, tool-use / harnesses, and LoRA/QLoRA/PEFT fine-tuning — optimizing for low latency, safe outputs, and learning-outcome alignment.
• Develop causal inference and reinforcement-learning pipelines (bandits, RLHF/RLAIF, policy optimization) that drive personalization and mastery velocity.
• Develop data science and data management workflows — transcripts, curriculum standards, chat interactions, feature stores, labeling, and experimentation frameworks.
• Contribute to fairness, transparency, and explainability through counterfactual analysis, model interpretability, and rigorous offline/online evaluation.
• Ship models to production with senior engineers — CI/CD, monitoring, model versioning, and drift detection.

Required Experience

• Education: PhD or MS (minimum) in Computer Science, Data Science, Machine Learning, Statistics, Applied Mathematics, or a closely related quantitative field, with a strong theoretical and mathematical foundation (probability, linear algebra, optimization).
• Industry Experience: 1–3 years of hands-on industry experience building and shipping ML/AI systems (substantive research internships and publications count toward depth).
• Data Science & Data Management: Strong foundations in statistical modeling, experimentation, feature engineering, and managing large, messy, real-world datasets.
• GenAI / LLMs: Hands-on with OpenAI (GPT), Google Gemini, Anthropic Claude, Mistral, Llama, or open-source LLM fine-tuning (LoRA/QLoRA, PEFT).
• Agentic Platforms: Experience building agents with first-party SDKs/APIs, or with frameworks such as LangGraph, AutoGen, CrewAI, or LlamaIndex, for multi-agent reasoning or workflow orchestration.
• Reinforcement Learning: Familiarity with Ray RLlib, Stable-Baselines3, CleanRL, TRL (HuggingFace), or custom policy-gradient implementations for bandit/RLHF/RLAIF pipelines.
• Causal AI: Skilled with DoWhy, CausalML, or SCM frameworks for counterfactuals, interventions, and causal inference applied to real-world (ideally education) data.
• Programming & Data: Strong Python; working knowledge of SQL and NoSQL; familiarity with React, Node.js, Java, and JSON a plus.
• MLOps: Exposure to deploying ML systems with CI/CD, monitoring (Prometheus/Grafana, Sentry, PostHog), model versioning, and drift detection.

Bonus points:

• Peer-reviewed publications in ML, RL, causal inference, or NLP.
• Passion for transforming education and closing equity gaps.
• Awareness of model safety, bias mitigation, and regulatory compliance (FERPA, COPPA, GDPR).