AI Engineer

Job Description

This defines the role, responsibilities, qualifications and evaluation criteria for the AI Engineer position at Fountain Forward. The hire will own the design and implementation of the company's next AI product, building agent-native platforms powered by multi-agent orchestration, Retrieval-Augmented Generation (RAG) and production-grade machine learning pipelines deployed on Databricks.

• AI Agent: An autonomous software component powered by a Large Language Model that reasons over a goal and executes actions through a defined tool-use interface.
• Orchestration Layer: The software boundary that coordinates multiple agents, tools, memory and control flow across an AI workflow, commonly implemented with frameworks such as LangGraph or LangChain.
• RAG (Retrieval-Augmented Generation): An architecture pattern where an LLM is grounded on external knowledge retrieved from a vector store or domain-specific catalog before generating a response.
• MLOps: The set of practices for building, deploying, monitoring and iterating machine learning models in production, typically covering experiment tracking, CI/CD and model governance.
• NLP (Natural Language Processing): The field of techniques for programmatically understanding, classifying and generating human language.
• LLM (Large Language Model): A foundation model trained on broad corpora and fine-tuned to follow instructions, used as the reasoning engine of an agent.

• Lead the technical build of Fountain Forward's first agent-native AI product, from prototype to production.
• Design and ship a robust orchestration layer that lets multi-agent workflows autonomously plan, retrieve, reason and act over enterprise data.
• Partner directly with product, data and client-facing teams to translate business problems into reliable AI capabilities.

• This is a builder role, not a research role: the work is measured by features shipped, uptime and business outcomes.
• The AI Engineer is expected to own the full stack of an AI feature, including agent design, tool interfaces, retrieval pipelines, model fine-tuning when relevant, and the data models that feed them.

• Design, build and maintain multi-agent workflows using LangGraph, LangChain and comparable orchestration frameworks.
• Implement typed, well-documented tool-use layers that expose backend capabilities (APIs, databases, ML models) to agents in a safe and auditable way.
• Build RAG systems over domain-specific catalogs, including chunking strategy, embedding model selection, vector store management and retrieval evaluation.
• Instrument agent runs with tracing, evaluation harnesses and regression suites to keep quality measurable as the system evolves.

• Design and implement the data models that back the AI product on Databricks, using a medallion (bronze, silver, gold) approach where applicable.
• Build scalable ETL pipelines that ingest from SharePoint, cloud storage, SaaS APIs and client systems into curated Delta tables consumable by agents and ML models.
• Apply data quality, schema validation and lineage controls so that every AI output can be traced back to a trustworthy source.
• Collaborate with data analytics to expose governed datasets for downstream reporting, including Power BI consumption layers.

• Develop, train and deploy machine learning models (classification, scoring, forecasting, NLP) that agents invoke as tools inside the orchestration layer.
• Wrap models behind clean, typed endpoints with predictable contracts so they can be composed by agents and other services.
• Maintain a disciplined MLOps workflow using MLflow for experiment tracking and CI/CD pipelines (Azure Pipelines, GitHub Actions or GitLab CI) for promotion to production.
• Monitor deployed models for drift, latency and cost, and iterate based on real user feedback.

• Work closely with the founding team to translate ambiguous product bets into concrete technical plans with realistic milestones.
• Write clean, reviewable code, document architectural decisions and participate actively in code reviews.
• Communicate trade-offs (accuracy vs. cost vs. latency vs. complexity) in language that both engineers and business stakeholders can act on.

• Bachelor's degree in Computer Science, Data Science, Artificial Intelligence, Software Engineering or a closely related engineering discipline.
• Coursework or demonstrated depth in at least two of: machine learning, data structures and algorithms, databases, cloud computing, MLOps.

• At least one (1) year of professional experience as a Data Engineer, AI Engineer, ML Engineer or equivalent applied role.
• Verifiable, hands-on experience shipping at least one AI or data product to production (not just notebooks or coursework).
• Experience building or contributing to agent-based systems with tool use, RAG or multi-step reasoning workflows.

• Advanced Python, including typed code, testing and packaging.
• Strong SQL and practical experience modeling data on Databricks (Delta Lake, Unity Catalog, Spark).
• Production experience with LangChain and LangGraph, or a clearly equivalent agent-orchestration framework.
• Experience building and deploying machine learning models that are consumed as tools, services or scheduled jobs.
• Comfortable working on a cloud platform (Azure preferred, AWS or GCP acceptable) and with Git-based workflows.

• Experience designing stateless, explainable scoring or decision engines whose outputs can be audited factor by factor.
• Experience building multi-tenant or multi-entity data frameworks with automatic schema detection and entity-specific business rules.
• Experience with MLflow, Azure Pipelines and Power BI as part of an end-to-end MLOps and reporting stack.
• Experience building full-stack services (e.g. FastAPI or Django backends, Next.js or similar frontends, containerized with Docker) to expose AI capabilities.
• Exposure to NLP techniques (embeddings, classification, topic modeling) and to at least one deep learning framework such as PyTorch, TensorFlow or Keras.
• Published research, conference presentations or open-source contributions related to AI, data science or agent systems.
• Prior internship or early-career recognition (awards, honors, distinguished-student status) that signals consistent high performance.

• Strong written and spoken English (B2 or higher). Additional languages are a plus.
• High ownership: comfort moving from a vague brief to a shipped feature with limited supervision.
• Clear, respectful communication with non-technical stakeholders.
• Genuine curiosity about the agentic AI landscape and a habit of tracking how the frontier is moving.

• Languages and data: Python (advanced), SQL, Spark.
• Agent stack: LangChain, LangGraph, RAG patterns, vector stores.
• Data and cloud: Databricks, Azure, Delta Lake, Git, CI/CD pipelines.
• ML and MLOps: scikit-learn or equivalent, MLflow, model deployment.

• Docker, FastAPI or Django, Next.js or a modern frontend framework.
• Power BI or Tableau for governed reporting on top of curated data.
• NLP tooling (NLTK, spaCy, Hugging Face), PyTorch, TensorFlow, Keras.
• Apache Airflow or equivalent orchestrator, DBT for transformations, Linux.
• Experience integrating external APIs (travel, hospitality, maps, enterprise SaaS).

• A founding-engineer seat on Fountain Forward's AI product, with meaningful influence on architecture and roadmap.
• A small, high-trust team that values shipped work, honest trade-offs and fast feedback loops.
• Direct exposure to senior leadership and to the business problems the AI product is solving.

• Competitive compensation aligned with the candidate's experience and the responsibilities of the role.
• Flexible remote or hybrid arrangement, with a strong preference for candidates in compatible Americas time zones.
• Budget for continuous learning, conferences and relevant tooling.

• Initial screening of CV, portfolio and production work (GitHub, papers, shipped products).
• Technical conversation focused on agent design, RAG and data modeling trade-offs the candidate has actually made.
• Practical exercise scoped to reflect real Fountain Forward problems (orchestration, tool design, data modeling on Databricks).
• Final interview with leadership to assess ownership, communication and cultural fit.

• Evidence of shipping, not just building.
• Clarity of thought when reasoning about AI system trade-offs.
• Depth in at least one of: agent orchestration, Databricks data modeling or applied ML.
• Alignment with Fountain Forward's values and working style.