Data Scientist

Data Scientist Remote, Anywhere in the US About Aperture Cloud At Aperture Cloud, we are trailblazing the future of autism therapy, making it more immediate, accessible, and effective for families everywhere. Our innovative virtual ABA therapy platform is thoughtfully designed by clinicians to recreate the focused, supportive environment of in-person therapy, complete with distraction-free features and interactive activities that enhance engagement and progress. Our team operates fully remote—meaning you’ll have the flexibility to work from the comfort of home. If you're ready to make a meaningful impact and join a team that's reshaping autism therapy, we’d love to hear from you! Why this role matters: As our Data Scientist, you will optimize the smart systems that pull real-time clinical context and turn it into safe, accurate, and highly relevant recommendations. By bridging the gap between cutting-edge AI capabilities and deep clinical expertise, you ensure our models are deeply rooted in real-world care and held to the highest quality standards. Your work ensures that our digital systems are a reliable, trusted partner for our clinical teams, allowing us to safely scale our platform and deliver life-changing autism therapy to families nationwide. Job Details W2 Employee Full-Time 100% Remote Job Requirements 4+ years of experience in applied data science, ML engineering, or AI engineering in a production environment Deep understanding of RAG architectures: retrieval systems, embedding models, vector databases (Pinecone, Weaviate, pgvector, or similar), chunking strategies, and context assembly Experience designing and running evaluation frameworks for AI systems — you've thought hard about how to measure quality in domains where ground truth is ambiguous Strong Python skills; experience with LLM orchestration frameworks (LangChain, LlamaIndex, or similar) Clinical NLP experience or healthcare AI background is strongly preferred — you understand why clinical data is different from general text and what that means for AI system design You think like an engineer and a scientist: you build systems that can be measured, iterated on, and trusted — not black boxes Strong written communication: you can explain RAG pipeline design to a clinician and explain clinical requirements to an engineer Genuine interest in the clinical domain — you want to understand Applied Behavior Analysis well enough to build AI that actually helps BCBAs do their jobs Nice to have: Experience with Amazon Bedrock, SageMaker, or AWS AI/ML services Familiarity with HIPAA-compliant data handling for AI training and inference pipelines Background in clinical NLP, behavioral health informatics, or ABA/autism research Experience with fine-tuning or RLHF — even if this role doesn't require it, understanding the tradeoffs informs better RAG design Exposure to LLM-as-judge evaluation patterns or multi-model evaluation pipelines What You’ll Do RAG Pipeline Design & Optimization Architect and continuously improve the RAG pipeline that retrieves client-specific clinical context — session notes, treatment plan goals, historical performance data — and injects it into inference-time prompts Design the retrieval layer: chunking strategies, embedding models, vector store configuration, and retrieval ranking — optimizing for clinical relevance, not just semantic similarity Build a context assembly system that selects and structures the most relevant clinical information for each model invocation, given token constraints and clinical priority Evaluate retrieval quality rigorously: build test sets, measure recall and precision, and iterate on the pipeline based on where retrieval fails Evaluation Framework Design Design evaluation frameworks that assess AI recommendation quality beyond standard NLP metrics — working with clinical stakeholders to define what 'good' means for each use case Build automated evaluation pipelines that can test AI outputs at scale: LLM-as-judge evaluators, human review workflows, and clinical validity checks Maintain evaluation datasets that reflect the real distribution of clinical scenarios the model encounters in production Report evaluation results in terms that clinical and product stakeholders can understand and act on Model Gap Analysis & Mitigation Systematically identify where foundation model capabilities fall short for Aperture Cloud's care model: what clinical reasoning the model gets wrong, what it hallucinates, what it doesn't know how to handle For each identified gap, recommend and implement the appropriate mitigation — improved retrieval, prompt engineering, output validation, or escalation to human review Stay current on foundation model capabilities and evaluate new models against our clinical requirements as they emerge Maintain a gap log and roadmap that gives product and clinical leadership visibility into current AI limitations and the plan to address them Production Monitoring & Quality Monitor production AI outputs