Lobello Sharon Ann

Senior Software Engineer and AI Evaluation Specialist with 6+ years of experience designing scalable backend systems, benchmarking large language models, building evaluation pipelines, and developing production-grade software across complex multi-repository environments. Strong expertise in Python, distributed systems, LLM evaluation methodologies, debugging large-scale applications, and creating robust benchmark frameworks for measuring reasoning, code quality, and software engineering performance of frontier AI models. Experienced in building automated evaluation harnesses, structured testing systems, synthetic benchmark generation pipelines, and failure analysis workflows used to assess model reliability, edge-case handling, and production readiness.

Proven ability to work across large codebases, lead technical initiatives, optimize CI/CD systems, and collaborate with research and engineering teams in fast-paced environments. Strong background in modern development workflows including Git, Docker, Kubernetes, cloud infrastructure, automated testing, and multi-language software environments. Adept at transforming ambiguous engineering problems into measurable evaluation systems and actionable technical insights.

Senior AI Evaluation Engineer at Scale AI (2024-01 – Present)

Led the design and implementation of advanced coding benchmark systems used to evaluate frontier large language models on real-world software engineering tasks including debugging, repository navigation, architecture reasoning, and production code generation.

• Designed benchmark suites covering: multi-file debugging, API integration tasks, dependency resolution, CI failure diagnosis, repository-wide refactoring, production-grade code implementation
• Built scalable Python-based evaluation pipelines capable of processing thousands of model-generated code submissions daily across multiple programming languages
• Developed automated grading harnesses for validating: correctness, runtime performance, security compliance, unit test coverage, code maintainability
• Implemented structured failure-analysis systems that identified recurring model weaknesses in: asynchronous programming, state management, edge-case handling, memory optimization, concurrency control
• Created synthetic evaluation datasets from real engineering scenarios gathered from large production repositories
• Collaborated with ML researchers and infrastructure teams to improve evaluation reliability and benchmark reproducibility
• Reduced benchmark execution latency by 38% through optimization of distributed task execution and caching strategies
• Built automated reporting dashboards for tracking model performance regressions and benchmark trends over time
• Conducted deep analysis of generated code quality and produced technical reports used to guide future model training improvements

Senior Backend Software Engineer at GitLab (2021-05 – 2023-12)

Worked on large-scale backend systems and repository management infrastructure supporting millions of developers worldwide. Contributed to platform reliability, CI/CD tooling, code review systems, and automated testing infrastructure.

• Maintained and optimized backend services handling high-volume repository operations and automated development workflows
• Designed internal tooling for large-scale repository analysis and automated code quality verification
• Built advanced CI/CD automation systems that improved deployment reliability and reduced failed builds across engineering teams
• Led debugging initiatives involving distributed microservices and complex production incidents
• Developed scalable APIs and asynchronous processing systems used by enterprise engineering customers
• Implemented comprehensive unit and integration testing pipelines that significantly improved platform stability
• Worked extensively within large, complex monorepos and multi-service architectures
• Participated in architecture reviews, technical planning, and engineering mentorship programs
• Improved observability and logging systems to accelerate root-cause analysis during production failures
• Contributed to developer tooling improvements for automated testing and merge validation workflows

Software Engineer – AI Infrastructure & Evaluation at Hugging Face (2020-02 – 2021-04)

Contributed to infrastructure and tooling supporting machine learning workflows, dataset evaluation systems, and model testing pipelines.

• Built automated evaluation systems for benchmarking NLP and code-generation models across standardized datasets
• Developed data ingestion pipelines for processing large-scale training and evaluation datasets
• Implemented tools for analyzing model-generated outputs and detecting hallucinations, syntax failures, and logic inconsistencies
• Created reproducible benchmark environments for testing model reliability across varying prompts and repository contexts
• Worked closely with research engineers to improve evaluation methodologies and metric consistency
• Developed Python utilities for dataset validation, transformation, and automated preprocessing
• Assisted in scaling distributed evaluation workflows running across cloud infrastructure environments
• Improved internal tooling for experiment tracking and benchmark result visualization

Backend Software Engineer at Red Hat (2018-07 – 2020-01)

Worked on enterprise backend systems and infrastructure automation projects within Linux and cloud-native environments.

• Developed scalable backend services supporting enterprise infrastructure management tools
• Built automation scripts and deployment systems for internal engineering operations
• Participated in large-scale debugging and performance optimization initiatives
• Improved system reliability through enhanced testing coverage and automated validation pipelines
• Contributed to containerized infrastructure projects and Kubernetes orchestration workflows
• Collaborated with cross-functional teams to implement production-ready engineering solutions
• Maintained secure coding practices and participated in code review processes across multiple teams

Master of Science (MSc) in Computer Science – University of Sheffield (2019 – 2021)

Bachelor of Science (BSc) in Software Engineering – University of Leeds (2015 – 2018)