Lead Data Scientist

Lead Data ScientistOctave's ETQ division is seeking a hands-on Data Scientist to build predictive models, implement Generative AI and Agentic AI features, and architect data-driven solutions for our document-based compliance management platform. This role requires a technical expert who can develop, deploy, and maintain ML systems in production environments.Responsibilities:Build and deploy Generative AI features using foundation models (AWS Bedrock, OpenAI, Anthropic Claude) and RAG architectures with vector databases for compliance document understandingDesign agentic AI systems that autonomously handle compliance workflows, document review, regulatory mapping, and multi-step reasoning tasksImplement comprehensive LLM evaluation frameworks with automated pipelines, custom metrics, benchmark datasets, and safety guardrails ensuring regulatory complianceBuild end-to-end MLOps pipelines for model training, deployment, monitoring, versioning, and automated retraining with drift detectionDevelop predictive models for compliance risk scoring, regulatory change impact, anomaly detection, and time-series forecastingWrite production-quality Python code for data processing, feature engineering, API development (FastAPI/Flask), and ETL/ELT workflowsLead A/B experiments and product analytics to measure AI feature impact and drive data-driven decision-makingCreate explainability frameworks (SHAP/LIME) and monitoring dashboards ensuring transparency and regulatory adherenceCollaborate with cross-functional teams to translate business needs into ML solutions and communicate insights to stakeholdersPython (5+ years): Production-level experience with Pandas, NumPy, scikit-learn, XGBoost, TensorFlow/PyTorch, Hugging Face Transformers, FastAPI/Flask, MLflow, and pytestSQL: Advanced proficiency with complex queries, window functions, and optimizationMachine Learning & NLP: Strong foundation in supervised/unsupervised learning, deep learning, document understanding, text classification, and semantic analysisGenerative AI & LLMs: Hands-on experience with foundation models (GPT, Claude, Llama), prompt engineering, RAG architectures, and vector databases (Pinecone, Weaviate, Chroma)MLOps & ModelOps: End-to-end experience with ML pipelines, experiment tracking (MLflow, W&B), model versioning, feature stores, drift detection, CI/CD for ML, and Docker containerizationLLM Evaluation: Experience with evaluation frameworks (RAGAS, DeepEval), custom metrics, benchmark datasets, and human-in-the-loop validationCloud & AWS: Experience with AWS services including SageMaker, Bedrock, S3, Lambda, EC2, and CloudWatchStatistics & Experimentation: Strong foundation in statistics, A/B testing, causal inference, and experimental designVisualization: Proficiency with Tableau, Power BI, or Python visualization librariesEducation / Qualifications:7+ years in data science, ML engineering, or related roles3+ years building NLP/generative AI applications and implementing MLOps in productionBachelor's or Master's degree in Data Science, Computer Science, Statistics, or related field (PhD preferred)Track record of deploying ML systems processing large-scale datasets with proper monitoring and governancePreferred Qualifications:Experience with agentic AI frameworks (LangGraph, LangChain, AutoGen, CrewAI)Knowledge of Life Sciences/regulated industries (FDA, EMA, ISO, GxP) and compliance management systemsFamiliarity with big data tools (Spark, Databricks, Snowflake), orchestration (Airflow, Kubeflow), and monitoring tools (Datadog, Prometheus)Experience with LLM fine-tuning, document processing libraries, multi-modal AI, or distributed trainingUnderstanding of ML governance, bias detection, model risk management, and data privacy regulations (GDPR, CCPA, HIPAA)Experience working in agile environments with JiraAWS ML certifications or similar credentialsKey Competencies:Strong communication skills explaining complex models to technical and non-technical audiencesAbility to work independently and collaboratively in fast-paced environmentsProven ability to convert POCs into production-grade solutionsUnderstanding of ethical AI and building trustworthy, explainable systems for regulated environments