SDE II

As an SDE II at LivePerson, your primary responsibilities will include managing the lifecycle of Large Language Models, collaborating with researchers for model fine-tuning, and developing scalable infrastructures for model serving. You will also design integration tools for applications and ensure smooth CI/CD deployment processes. Your role is pivotal in enhancing our Conversational AI capabilities.

To qualify for the SDE II role at LivePerson, you should have a Bachelor's or Master's degree in Computer Science or a related field, along with at least 3 years of software development experience. Strong skills in Java or Python, familiarity with cloud platforms like GCP, and containerization technologies such as Docker and Kubernetes are essential. Understanding NLP concepts and experience with Large Language Models are preferred.

As an SDE II at LivePerson, familiarity with programming languages such as Java or Python, along with experience in cloud environments and containerization technologies, is crucial. Additionally, experience with frameworks like PyTorch and HuggingFace Transformers will be highly beneficial for working on Large Language Models.

LivePerson is committed to employee development and offers comprehensive support for SDE IIs. We provide tuition reimbursement for further education, access to AI learning resources, and a culture that encourages continuous growth and innovation. You will have ample opportunities to collaborate with industry experts and stay updated on the latest advancements in NLP and AI.

At LivePerson, you can expect an inclusive and collaborative work culture. We celebrate diversity and empower our teams to forge impactful conversations globally. We highly value innovation and encourage all employees to share ideas, ask questions, and contribute to a shared goal of improving customer experience through technology.

When designing a model lifecycle management system, consider aspects like version control, documentation, and performance tracking. Outline how you would implement change tracking and ensure collaboration among teams during the development process.

In model fine-tuning, I begin by understanding the specific requirements of the task or domain. I typically utilize metrics such as accuracy, precision, recall, and F1-score to evaluate the performance, making adjustments based on data feedback and experimental results.

To ensure high availability and low latency, I would implement load balancing, monitor performance in real-time, and optimize resource allocation. Regular testing under different scenarios can also help to proactively identify bottlenecks and ensure stability.

Sure! I once faced a challenge with the deployment of a new model where it wasn’t performing as expected in production. I facilitated a team brainstorming session, gathered logs to analyze the failure points, and collaboratively we identified a misconfiguration in the environment that we corrected, improving overall performance.

I regularly read research papers, attend industry conferences, and participate in online courses on platforms like Coursera. Keeping an eye on innovations from major conferences like NeurIPS or ACL also helps me gain insights into the latest trends and techniques.

CI/CD is vital for ensuring rapid and reliable deployment of code changes. I embrace automation in testing, integration, and deployment, which reduces the chance of human error and speeds up the release process while maintaining model accuracy and reliability.

I have used Docker to create containerized applications that streamline deployment processes. With Kubernetes, I have orchestrated containers for scaling applications seamlessly, ensuring efficient resource utilization and high availability.

Successful collaboration requires effective communication and regular updates. I establish clear goals with stakeholders and schedule periodic check-ins to ensure alignment, while inviting feedback throughout the development cycle.

One challenge is ensuring compatibility with existing systems. To address this, I would assess the technical requirements early and design integration libraries that facilitate seamless interaction between the LLMs and the application infrastructure, incorporating feedback from application teams.

Documentation is crucial in managing Large Language Models as it enhances understanding of the models, facilitates better collaboration, and serves as a reference for future improvements. Maintaining comprehensive documentation on configurations, performance, and any changes is essential to ensure consistency.