• Design and implement training data generation pipelines, including synthetic data generation. • Build data labeling and annotation workflows with quality validation loops. • Convert enterprise data into formats suitable for model training (instruction-tuning pairs, embeddings). • Implement active learning strategies to identify high-value training examples. • Collaborate with domain experts to validate training data quality and relevance. • Build automated data quality checks: coverage, balance, consistency. • Design training data versioning and lineage tracking. • Analyze model evaluation results to identify training data gaps.

• Rethink the PM job when most users are agents. • Talk to customers across the full spectrum. • Own the problem statement and requirements behind every agent tooling bet. • Decide what gets built, what gets deferred, and what gets cut. • Define how each launch is measured before it ships. • Keep engineering, design, and leadership aligned.

• Manage and evolve the end-to-end architectural blueprint for cloud-native applications at scale — covering compute, data, networking, security, and observability layers on AWS. • Define and promote agentic AI architecture patterns, including multi-agent systems, tool orchestration, memory and retrieval strategies, LLM routing, guardrails, and feedback loops. • Translate these patterns into relevant standards and reusable reference architectures. • Establish and govern DevOps standards across the organization: CI/CD pipeline design, GitOps practices, deployment strategies (blue/green, canary, progressive delivery), environment parity, and release engineering. • Design and oversee platform infrastructure. This includes ECS, Kubernetes clusters (EKS), service mesh, API gateway strategy, secrets management, IAM/RBAC governance, and data stores like DynamoDB, Aurora RDS, Postgres, Elastic/OpenSearch. Additionally, it involves multi-account/multi-region AWS topology. • Lead architecture reviews and provide binding technical decisions on high-risk changes — balancing velocity, reliability, cost, and security. • Define SLOs/SLIs/error budgets, logging standards, distributed tracing, capacity planning, and incident response frameworks. • Introduce new technologies through structured PoC and risk assessment processes; build internal communities of practice around architectural patterns. • Produce architecture decision records (ADRs), system context diagrams, threat models, and runbooks that serve as the source of truth for platform design. • Be a technical advisor to Product and Partners — translating architectural constraints and capabilities into clear strategic options. • Mentor staff and senior engineers on architectural thinking, systems design, and engineering leadership; conduct design reviews that raise the technical bar across teams.

• You will architect and implement the core AI pipeline that powers Accel's test creation system. • Work closely with the founder to design and build an AI-powered content generation system from the ground up. • You'll contribute to meaningful parts of the product end-to-end from how the system ingests and understands source material, to how it produces and validates outputs, to how instructors interact with and review what the system generates. • On the engineering side, you'll build and iterate on LLM-driven pipelines, work with retrieval and embedding techniques to ground outputs in real source material and develop backend services and APIs that tie everything together. • Beyond pure coding, you'll be expected to think about output quality and building evaluation steps, catching failure modes, and improving the system based on real instructor feedback. • You'll research new tools and techniques as the AI space evolves and bring relevant ideas directly into the product. • This is a generalist role at an early-stage product where you'll wear multiple hats, work with ambiguity, and have direct input into how things are built.