Machine Learning Engineer - LLMs and Agentic

About Oversight Oversight is the world s leading provider of AI-based spend management and risk mitigation solutions for large enterprises. Based in Atlanta, GA, Oversight works with many of the world s most innovative companies and government agencies to digitally transform their spend audit and financial control processes. Oversight s AI-powered platform works across our customers financial systems to continuously monitor and analyze all spend transactions for fraud, waste, and misuse. With a consolidated, consistent view of risk across their enterprise, customers can prevent financial loss and optimize spend while strengthening the controls that improve compliance. LearnMore. PositionOverview: We are seeking a skilled and forward-looking ML Engineer with experience in Large Language Models (LLMs), generative AI, and agentic architectures to join our growing R&D and Applied AI team. This role is critical in helping Oversight deliver the next generation of agentic AI systems for enterprise spend management and risk controls. The ideal candidate has a strong foundation in machine learning, modern deep learning frameworks, and data pipelines, coupled with hands-on experience experimenting with LLMs, small language models (SLMs), multi-agent frameworks, and retrieval-augmented generation (RAG). You will work closely with AI/ML researchers, data engineers, and product teams to design, implement, and optimize models that power autonomous exception resolution, anomaly detection, and explainable insights. This is a hands-on engineering role where you will not only build and scale ML systems but also actively contribute to cutting-edge applied research in agentic AI. Core ML/LLM Engineering • Contribute to the design, training, fine-tuning, and deployment of ML/LLM models for production. • Implement RAG pipelines using vector databases. • Work with frameworks like LangChain, LangGraph, MCP to prototype and optimize multi-agent workflows. • Develop prompt engineering, optimization, and safety techniques for agentic LLM interactions. • Integrate memory, evidence packs, and explainability modules into agentic pipelines. • Work hands-on with multiple LLM ecosystems: • OpenAI GPT models (GPT-4, GPT-4o, fine-tuned GPTs). • Anthropic Claude (Claude 2/3 for reasoning and safety-aligned workflows). • Google Gemini (multimodal reasoning, advanced RAG integration). • Meta LLaMA (fine-tuned/custom models for domain-specific tasks). Data & Infrastructure • Collaborate with Data Engineering to build and maintain real-time and batch data pipelines that serve ML/LLM workloads. • Conduct feature engineering, preprocessing, and embeddings generation for structured and unstructured data. • Implement model monitoring, drift detection, and retraining pipelines. • Leverage cloud ML platforms (AWS Sagemaker, Databricks ML) for experimentation and scaling. Research & Applied Innovation • Explore and evaluate emerging LLM/SLM architectures and agent orchestration patterns. • Experiment with generative AI and multimodal models to extend capabilities beyond text (images, structured financial data). • Collaborate with R&D to prototype autonomous resolution agents, anomaly detection models, and reasoning engines. • Translate research prototypes into production-ready components. Collaboration & Delivery • Work cross-functionally with R&D, Data Science, Product, and Engineering to deliver business-aligned AI features. • Participate in design reviews, architecture discussions, and model evaluations. • Document processes, experiments, and results effectively for knowledge sharing. • Mentor junior engineers and contribute to ML engineering best practices. Education,ExperienceandSkills Required • Bachelor s or Master s degree in Computer Science, Data Science, MachineLearning, or related field. • 3+ years of experience building and deploying ML systems. • Proficiency in Python and libraries such as PyTorch, TensorFlow, Scikit-Learn, Hugging Face Transformers. • Hands-on experience with LLMs/SLMs (fine-tuning, prompt design, inference optimization). • Demonstrated experience with at least two of the following ecosystems: • OpenAI GPT models (chat, assistants, fine-tuning). • Anthropic Claude (safety-first AI for reasoning and summarization). • Google Gemini (multimodal reasoning, enterprise-scale APIs). • Meta LLaMA (open-source, fine-tuned models). • Familiarity with vector databases, embeddings, and RAG pipelines. • Ability to work with structured and unstructured data at scale. • Knowledge of SQL and distributed data frameworks (Spark, Ray). • Strong understanding of ML lifecycle: data prep, training, evaluation, deployment, monitoring. Preferred Qualifications • Experience with agentic frameworks (LangChain, LangGraph, MCP, AutoGen). • Knowledge of AI safety, guardrails, and explainability techniques. • Hands-on experience deploying ML/LLM solutions in cloud environments (AWS, GCP, Azure). • Ex...