Citi
Citi
Citi je globalno bančno in finančno podjetje, ki zagotavlja storitve potrošnikom, korporacijam, vladam in institucijam. Njegova ponudba vključuje bančništvo za prebivalstvo, kreditne kartice, upravljanje premoženja, poslovno bančništvo in investicijsko bančništvo, podprto s poslovanjem v več kot 100 državah. Kot velika mednarodna finančna organizacija Citi združuje ekipe, ki delujejo na področjih bančništva, finančnih storitev in fintecha.

Senior Lead Python AI and Machine Learning Engineer

Lead the architecture and delivery of scalable generative AI and predictive machine learning systems for Citi’s US retail banking fraud prevention. Drive MLOps practices and mentor engineering teams.

Opis

  • Guide the development and launch of advanced AI and machine learning products.
  • Architect production generative AI solutions, from retrieval-augmented generation and autonomous agents to semantic search and multimodal workflows.
  • Tune Google Gemini, Anthropic Claude, and open-source language models for speed, cost, and quality.
  • Develop prompt engineering frameworks, model fine-tuning pipelines, evaluation benchmarks, and safety controls.
  • Set predictive ML architecture for anomaly detection, graph neural networks, and supervised and unsupervised classification.
  • Build dependable, fast, horizontally scalable Python services and APIs.
  • Connect vector databases, distributed caches, and event streaming systems.
  • Establish MLOps and LLMOps workflows for CI/CD, automated retraining, observability, and drift detection.
  • Create AI capabilities to identify identity theft, account takeovers, synthetic identities, card-not-present fraud, and unusual transactions.
  • Apply generative AI to streamline fraud triage and summarize investigations from unstructured information.
  • Maintain compliance with US banking regulations.
  • Coach data scientists and software engineers in the Chennai engineering hub.
  • Set the technical roadmap, assess AI frameworks, run proofs of concept, and present architecture proposals to global engineering leaders.
  • Improve fraud detection, false-positive rates, inference speed, code quality, test coverage, and investigator efficiency.

Zahteve

  • Hold a bachelor’s or master’s degree in computer science, data science, artificial intelligence, applied mathematics, or a related quantitative discipline.
  • Bring 10–12 years of professional software engineering experience, including development of production AI and ML systems.
  • Demonstrate advanced Python 3.x skills, including multiprocessing or asyncio, design patterns, profiling, packaging, and clean coding practices.
  • Have substantial applied experience with PyTorch, TensorFlow, Scikit-learn, XGBoost, and LightGBM.
  • Show a record of building production generative AI applications with tools such as LangChain, LlamaIndex, CrewAI, or AutoGen.
  • Have practical knowledge of embedding models, dense retrieval, reranking, and vector indexing methods.
  • Bring hands-on experience with AWS, GCP, or Azure, plus Docker, Kubernetes, and PySpark or Ray.
  • Have delivered end-to-end ML lifecycles, from data validation through real-time inference monitoring.
  • Have integrated Google Gemini and Anthropic Claude APIs, including tool or function calling, context caching, and prompt optimization.
  • Experience with fraud rule engines, anomaly detection platforms, or predictive risk models is preferred.
  • Familiarity with transaction monitoring, behavioral biometrics, and fraud consortium datasets is preferred.
  • Be highly proficient in Python, SQL, and Bash.
  • Have experience with vector databases, backend APIs, data streaming, MLOps, DevOps, explainable AI, and masking personally identifiable information.
  • Be able to meet US banking regulatory requirements, including FCRA, GLBA, SR 11-7, fair lending, and explainable AI standards.

Ugodnosti

  • Hybrid work arrangement.
  • Equal opportunity employer.
  • Reasonable accommodations available for people with disabilities.

Sorodna delovna mesta