Senior Agentic AI and Harness Engineer

HS Financial Technologies; Województwo pomorskie; Hybride; PLN 180 000 - 280 000; Plein temps

Avantages offerts par ce poste

LuxMed Premium Package

Life insurance

Employee Referral Program

Remote/hybrid working opportunities

100% funding for certifications and CP

Technical workshops and soft skills

Résumé du poste

HS Financial Technologies is seeking a senior software engineer to design, build, and operate production AI agents for financial operations and enterprise workflows.

You will own the agent runtime, orchestration, memory, tool execution, and safety controls, with a focus on secure, auditable, and scalable production pipelines. Strong Python, AWS, and LLM experience are required, along with cross-functional collaboration.

Qualifications

• Bachelor’s degree in a relevant technical field or equivalent practical experience.
• 5+ years of software engineering experience, including at least 1 year delivering LLM or agent-based applications in production.
• Strong Python, API design, asynchronous processing, automated testing, source control, and AWS engineering skills; experience with a managed model platform such as Amazon Bedrock or an equivalent service.
• Hands‑on experience implementing agent loops, context assembly, retrieval, memory, tool selection and limits, durable state, checkpoints, retries, and idempotency.
• Ability to build regression evaluations for outcomes, evidence, and tool paths; enforce least‑privilege tool access and prompt‑injection defenses; and separate model reasoning from deterministic controls.
• Strong troubleshooting, analytical, communication, and cross‑functional collaboration skills.

Responsabilités

• Runtime: Design, deploy, and support reliable production runtimes for AI agents serving financial operations and enterprise workflows.
• Orchestration: Build and extend agent loops, including context assembly, retrieval, memory, tool selection, execution limits, workflow state, and response validation.
• Controls: Enforce authorization, human approvals, policy checks, validation, checkpoints, idempotency, retries, and exception handling throughout agent workflows.
• Enterprise integration: Build validated, auditable tool and system interfaces with least-privilege permissions, prompt-injection defenses, and controls outside the model that isolate reasoning from deterministic execution.
• Evaluation: Build repeatable tests and regression suites for agent outcomes, supporting evidence, and tool‑execution paths; diagnose quality, safety, failures, latency, and cost.
• Observability: Monitor and troubleshoot production agents, maintain audit evidence, and turn operational findings into reusable engineering standards.

Description du poste

The AI Engineering team builds secure, reliable, and reusable AI capabilities for HS Financial Technologies’ business and technology platforms, partnering across engineering, security, data, and operations to move solutions into production.

We integrate AI into existing workflows and platforms, focusing on measurable outcomes, production discipline, responsible AI, security, human oversight, and evidence-based evaluation.

Role Overview

You will design, build, and operate production AI agents for financial operations and enterprise workflows. In this role, the AI harness is the production layer around the model: the agent runtime, orchestration, controls, evaluation, and observability. You will own this layer, including retrieval, memory, tool execution, workflow state, human approvals, and safety controls.

You will bring strong software engineering skills, hands-on production LLM experience, sound architecture judgment, effective troubleshooting, and cross-functional collaboration. You will also establish reusable patterns for HSFT engineering teams.

Key Responsibilities

• Runtime: Design, deploy, and support reliable production runtimes for AI agents serving financial operations and enterprise workflows.
• Orchestration: Build and extend agent loops, including context assembly, retrieval, memory, tool selection, execution limits, workflow state, and response validation.
• Controls: Enforce authorization, human approvals, policy checks, validation, checkpoints, idempotency, retries, and exception handling throughout agent workflows.
• Enterprise integration: Build validated, auditable tool and system interfaces with least-privilege permissions, prompt-injection defenses, and controls outside the model that isolate reasoning from deterministic execution.
• Evaluation: Build repeatable tests and regression suites for agent outcomes, supporting evidence, and tool‑execution paths; diagnose quality, safety, failures, latency, and cost.
• Observability: Monitor and troubleshoot production agents, maintain audit evidence, and turn operational findings into reusable engineering standards.

Required Qualifications

• Bachelor’s degree in a relevant technical field or equivalent practical experience.
• 5+ years of software engineering experience, including at least 1 year delivering LLM or agent-based applications in production.
• Strong Python, API design, asynchronous processing, automated testing, source control, and AWS engineering skills; experience with a managed model platform such as Amazon Bedrock or an equivalent service.
• Hands‑on experience implementing agent loops, context assembly, retrieval, memory, tool selection and limits, durable state, checkpoints, retries, and idempotency.
• Ability to build regression evaluations for outcomes, evidence, and tool paths; enforce least‑privilege tool access and prompt‑injection defenses; and separate model reasoning from deterministic controls.
• Strong troubleshooting, analytical, communication, and cross‑functional collaboration skills.

Preferred Experience

• Experience in financial services or another regulated or highly controlled environment.
• Experience with model or tool gateways and interoperability protocols, such as MCP, plus multi‑agent systems and durable workflow engines.
• Knowledge of secure tool execution, containers, infrastructure as code, CI/CD, and observability platforms.
• Familiarity with responsible AI, governance, model risk management, and regulatory controls.
• Experience building shared AI platforms, mentoring engineers, or establishing reusable engineering standards.
• Relevant certifications in AWS, cloud engineering, AI, security, or data.

We’re proud to offer a benefits package and workplace culture designed to support your health, career ambitions and life beyond work.

• LuxMed Premium Package
• Life insurance and travel insurance
• Employee Referral Program
• Remote/hybrid working opportunities, or from the office
• 100% funding for certifications and postgraduate studies accredited by ACCA, CIMA and IOB
• Technical workshops and soft skills training