Mid-Level Data Scientist

Quick Facts

• Role: Mid-Level Data Scientist (LLM/Agentic Solutions)

Description

Develop and industrialize LLM-based agentic solutions for the Investment Banking division, with high-impact use cases in document analysis, credit, and finance. Own the design of data extraction pipelines, evaluation and observability frameworks, and deploy models on GPU infrastructure using robust testing and DevOps practices. Collaborate with business teams to drive adoption while respecting LLM non-determinism and compliance constraints.

Responsibilities

• Design and develop LLM-based agentic solutions to automate complex business processes

• Partner with business teams to identify needs and drive AI adoption

• Build data extraction pipelines and implement evaluation and observability for AI agents

• Deploy and industrialize models (LLMs, GPUs), ensuring quality through testing and DevOps best practices

• Take full ownership of delivering AI features

• Implement safety rails and clearly communicate constraints and trade-offs

• Prototype quickly (PoCs), validate feasibility with data, and avoid technical debt

• Mentor peers and work effectively with architects and stakeholders

Requirements

• Strong Python with OOP and enterprise software design patterns

• Experience building asynchronous microservices/endpoints with FastAPI

• Hands-on pytest testing for ML/LLM pipelines (unit, functional, integration)

• Bash scripting for automation, data handling, and container interaction

• Demonstrated understanding of LLM architectures via portfolio

• Experience with multi-step, stateful agentic graphs using LangGraph or Google ADK

• Programmatic optimization of LLM pipelines using DSPy

• Experience deploying and running high-throughput open-weight models on GPU infrastructure using runtimes like vLLM; CUDA and fine-tuning are a plus

• Experience with Jenkins for automated build, test, and release pipelines

• Experience with tracing, drift detection, and automated evaluations (e.g., Arize Phoenix)

• Ability to ship stable, secure, compliant code for regulated banking infrastructure

• General stack exposure: Google ADK, LangGraph, DSPy, vLLM, FastAPI, Python, pytest, Arize Phoenix, Kubernetes, Jenkins, GPUs

• Nice to have: Corporate Institutional Banking project experience

Benefits

• Training, certification exam, and postgraduate studies co-financing

• Permanent contract or B2B cooperation

• Opportunity to influence the company direction; openness to your ideas

• Excellent working atmosphere and integration events