Senior Software Engineer I, Machine Learning

As a Senior Software Engineer I, Machine Learning at Cisco ThousandEyes, your primary responsibilities will include designing, implementing, and maintaining large-scale AI/ML pipelines that power our anomaly detection systems. You'll be involved in training and tuning machine learning models, particularly focusing on algorithms like Isolation Forests and LSTM-based frameworks. Furthermore, your role will encompass developing sophisticated evaluation frameworks to assess model performance and optimizing stream processing solutions using technologies like Flink and Kafka.

To be considered for the Senior Software Engineer I, Machine Learning position at Cisco ThousandEyes, candidates should have 3-5 years of software development experience, including at least two internships with direct experience in building machine learning models. An MS or PhD in a relevant field is preferred. Applicants should be proficient in Python and familiar with machine learning frameworks such as SKLearn, XGBoost, PyTorch, or TensorFlow, and able to translate concepts into effective algorithms.

When you apply for the Senior Software Engineer I, Machine Learning role at Cisco ThousandEyes, your application will be reviewed by our recruiting team. The application window closes on April 7, 2025, but please note that we may remove the posting earlier if the position is filled. We encourage applicants to submit their resumes as soon as possible to ensure consideration. After reviewing applications, selected candidates will be invited for interviews where they can showcase their skills and experiences.

In the Senior Software Engineer I, Machine Learning role at Cisco ThousandEyes, you'll work extensively with machine learning frameworks such as PyTorch, TensorFlow, and SKLearn. You will also utilize stream processing technologies like Flink and Kafka, offering numerous opportunities to innovate and optimize real-time data handling processes. Your experience with large-scale processing systems will be invaluable as you tackle diverse datasets.

Cisco ThousandEyes offers a comprehensive benefits package for the Senior Software Engineer I, Machine Learning position, including medical, dental, and vision insurance, a 401(k) plan with matching contributions, generous vacation policies, and flexible time off. Employees also enjoy paid holidays, sick leave, and unique programs tailored to their personal and community needs, ensuring a well-rounded and supportive work environment.

When designing machine learning models for anomaly detection, I start with a clear understanding of the data and the types of anomalies I want to detect. I analyze the dataset to identify patterns and trends before selecting the appropriate algorithms, such as Isolation Forests or LSTMs. It’s crucial to create a pipeline that includes preprocessing steps, model training, evaluation metrics, and deployment strategies to ensure the model's effectiveness in real-time scenarios.

In my previous roles, I have implemented various anomaly detection algorithms, including Isolation Forests and Variational Autoencoders. I focus on understanding each algorithm's strengths and weaknesses, adapting them as needed based on the data characteristics. For example, using LSTM networks has been beneficial in time-series data where temporal relationships play a key role in identifying anomalies.

Python is fundamental to my development process due to its simplicity and extensive libraries for machine learning, such as TensorFlow and PyTorch. I leverage these frameworks to streamline the building and evaluating of models while ensuring scalability. Python's versatility also allows easy integration with data processing tools like Flink and Kafka, enhancing my ability to create robust streaming applications.

I evaluate the performance of my machine learning models through a combination of metrics depending on the objectives, such as precision, recall, F1 score, and AUC-ROC curves. I also utilize validation sets and cross-validation techniques to ensure that models generalize well to unseen data. Continuous monitoring post-deployment helps identify performance drifts, allowing for timely adjustments.

Collaboration within a team of engineers is paramount to successfully executing machine learning projects. I prioritize open communication and actively share insights and feedback during meetings. Utilizing tools like Git for version control ensures a smooth workflow, and I’m a proponent of pair programming to foster knowledge sharing and maintain code quality among team members.

I stay current with developments in machine learning by subscribing to reputable journals, attending conferences, and participating in community forums and discussion groups. Engaging with online courses and workshops on platforms like Coursera or edX also helps me deepen my understanding of emerging trends and technologies that I can apply to my work.

One significant challenge I faced was dealing with unstructured data that required extensive preprocessing before training could begin. I addressed this by developing a robust data pipeline that automated the cleaning and feature extraction processes, allowing for a smoother transition to training. Learning to adapt and iterate quickly in response to these challenges is a key part of my approach.

In a past role, I led a project to develop a real-time anomaly detection system for network traffic analysis. By collaborating closely with cross-functional teams and leveraging LSTM networks, we achieved a significant reduction in false positives, which improved system response times. This project not only honed my technical skills but also enhanced my leadership and project management experience.

Large-scale machine learning refers to techniques and processes that manage massive datasets and complex models effectively. It involves optimizing algorithms and infrastructures to deploy models across distributed systems, ensuring that they can handle real-time data processing needs, which is particularly crucial in environments like Cisco ThousandEyes where large amounts of telemetry data are analyzed.

In a fast-paced development environment, I prioritize tasks based on their impact and urgency. Utilizing frameworks like Agile helps me break down large projects into manageable sprints, allowing for regular assessments of progress and adjustments to priorities. I also maintain open communication with my team to align on goals and ensure collaboration on high-impact tasks.