Specialist - Tech Development

Job description

Position Summary The Tech Development Lead is a Grade 8 individual contributor who supports design, build, deployment, and operates reliable data pipelines and AI-enabled solutions. The role applies to technical judgment across data engineering, GenAI, and agent engineering, with accountability for production reliability, observability, trusted data access, prompt effectiveness, and responsible operation in a regulated environment. The role owns complex components of the data and AI landscape end-to-end, makes day-to-day technical decisions with limited oversight, and helps shape practical engineering direction. It partners with business, product, platform, architecture, security, governance, and engineering teams to translate ambiguous needs and user feedback into scalable, maintainable solutions while mentoring colleagues and strengthening team capability. Why This Role Exists Own data pipelines, datasets, and supporting interfaces end-to-end for analytics, machine learning, and GenAI use cases. Build and maintain AI agents and LLM-powered capabilities that support trusted, efficient business workflows. Establish agent reliability, observability, guardrails, and escalation controls for stable and compliant production operation. Use AI-assisted engineering as a core working practice while applying strong judgment to generated code, tests, documentation, recommendations, and downstream risk implications. Key Responsibilities 1. Data Pipeline & Platform Engineering Independently design, build, deploy, and operate batch and streaming data pipelines that meet defined expectations for freshness, correctness, reliability, performance, and cost. Own data models and datasets across ingestion, transformation, storage, and serving, using modern warehouse or Lakehouse technologies. Monitor source and platform changes, including ServiceNow data exposed through Databricks views, and address inconsistencies that could affect downstream solutions or agent reliability. Make informed technical decisions on pipeline, interface, data-model, and automation changes, escalating only where architectural, compliance, or business-risk thresholds require broader approval. Control and monitor approved data access points across structured and unstructured sources, cloud platforms, documentation, and authoritative business repositories. 2. AI Agent & GenAI Engineering Build, deploy, and enhance machine learning and GenAI solutions aligned with validated user requirements. Develop and maintain AI agents using Python, Azure services, APIs, LLM frameworks, retrieval patterns, and appropriate interface technologies. Use AI coding assistants and LLM tooling to plan, scaffold, refactor, test, document, and debug code, while validating outputs before adoption. Create agent-based automations for data-quality investigation, failure triage, documentation, schema and lineage analysis, dataset discovery, and routine remediation. Test, refine, and govern prompts; identify new prompts aligned with evolving business needs and approved solution scope. Incorporate user feedback, agent metrics, and production outcomes into a closed-loop improvement process for prompts, models, and agent behavior. 3. Reliability, Observability & Trust Controls Monitor system health, trace logs, response quality, hallucination indicators, data drift, failures, and operational trends. Define and maintain practical detection thresholds, alerts, dashboards, and escalation paths for production issues. Instrument pipelines and agents, respond to incidents, perform root-cause analysis, and drive corrective actions for reliability, quality, security, performance, and cost. Validate that agent responses rely on approved authoritative data sources and map to defined domain controls, identifiers, measures, or calculation logic where applicable. Implement refusal or human-escalation behavior when a request is outside approved scope, lacks authoritative evidence, or requires human judgment. Partner with AI platform and technology teams to improve guardrails, usability, response clarity, and front-end experience without changing approved business logic. 4. Quality, Governance & Documentation Implement data-quality and testing controls such as freshness checks, contract tests, anomaly detection, test automation, and clear alerting paths. Ensure changes align with applicable architecture, security, privacy, platform, model-governance, and responsible-AI standards. Maintain technical documentation, operating procedures, training materials, data lineage, governance records, and service-landscape knowledge. Act as an effective first-line tester for changes and enhancements, with clear evidence of validation and traceability. Influence architecture and tooling decisions, balancing delivery speed with maintainability, supportability, scalability, and compliance. 5. Business Partnership & Capability Building Translate business needs, ambiguous requirements, production findings, and user feedback into prioritized, deliverable increments. Consult business and technology stakeholders on proposed changes and communicate technical implications clearly to non-technical audiences. Partner with product owners, risk or control specialists, data and analytics teams, and platform teams to deliver solutions that preserve approved business logic. Train end users on effective prompt usage, solution boundaries, escalation paths, and good practices. Mentor engineers and colleagues through pairing, code review, onboarding, troubleshooting support, and practical knowledge sharing; act as a go-to technical advisor within the team. 6. Development Capability (ability to fulfil at least 1 of the following specialisms) Front-end development – ability to design, build, test and maintain accessible, responsive user interfaces using React, JavaScript/TypeScript, HTML5 and CSS, or comparable GSK-supported frameworks. Ability to translate user interviews, observation and usability feedback into intuitive, secure user experiences through prototyping, iterative development and validation. Data engineering – ability to design and operate secure, reliable ingestion from ADLS, Snowflake or other GSK-supported components into Databricks and the appropriate Code Orange foundation, trusted or unified layer. Ability to validate, transform, clean, document and model data through maintainable queries and jobs, producing governed tables or payloads for analytics, data science, AI/ML and LLM solutions, including structured outputs consumable by APIs and front-end applications. Data science – ability to design, build, evaluate, deploy and support AI/ML models for text, image or numerical analysis. Select reproducible statistical or machine-learning approaches appropriate to the business need and data, define meaningful performance measures, and validate model quality, limitations, robustness and ongoing performance. AI engineering – ability to design, version, test and optimise prompts, retrieval and agent workflows; integrate approved LLMs through APIs; and select models and reasoning settings against defined quality, latency, security and cost criteria. Ability to use AIGA components for evaluation, registration, deployment and telemetry, including Weights & Biases Weave and OpenTelemetry, to trace lineage, usage, cost, performance and explainability. Full-stack development – ability to design, implement, secure and test end-to-end applications and API-based data interchange across user interface, service and data layers. Demonstrated capability in at least two other listed specialisms, including integration, deployment, observability and operational support within the GSK technology stack. Required Qualifications & Experience Bachelor's degree in computer science, engineering, data science, data engineering, or a related field, or equivalent practical experience. Typically 4- years of relevant hands-on experience building, operating, and improving productio