Sr. Applied Scientist

Job description

DESCRIPCIÓN

The PeopleInsight team is looking for a Machine Learning (ML) Engineer to help our ML group develop and deploy machine learning solutions for various teams within Worldwide Operations. The ML team builds scalable end-to-end impressive production machine learning solutions for World Wide Operations. Our solutions include; predicting the location and time of critical events, NLP/NLU tasks, developing and serving recommender systems, and running simulations that are used for long term strategic planning in World Wide Operations. ML models are typically deployed as components of larger software projects so we work collaboratively with teams consisting of product managers, UX designers, data engineers and operations leaders. We build and deploy models following high software standards emphasizing speed and scale, and leveraging world class AWS infrastructure. The team constantly challenges assumptions and works to clarify how models perform, and how they impact the user experience. The ML team also invests heavily in R&D, as the field evolves, we will remain in touch with the state-of-the-art. This is an exciting area of Amazon to gain real world ML experience and insight!

We are now recruiting an exceptional Machine Learning Engineer for Worldwide Operations.

Success in this role will require:

Strong Foundations
• The ML engineer will write efficient and prudent programs in Python and demonstrate attention to detail when debugging. A successful candidate will anticipate software bugs as well and write unit tests to detect them before serving models to production.
• The ML engineer should understand the ML algorithms being implemented beyond the use of packages, but at a mathematical level. This includes the ability to approximate program runtime and space complexity.

Data Handling Skills
• ML engineers will help develop the data acquisition pipelines and data labeling workflows with embedded SQL, pySpark, or Athena.
• The ML engineer should be able to deal with various file types, data formats, and data structures

ML Model Training Expertise
• Experience training traditional ML models: trees, regressions, nearest neighbors; as well as Deep Learning models
• Experience with specialized deep learning methods; RNNs and LSTMs for NLP/NLU, and CNNs for vision tasks
• Patiently conducting error analysis on a baseline ML models, and deliberate and clear-eyed tuning to improve model performance. The ML engineer should consider constraints in the data pipeline and UI phases when evaluating models

Model Deployment and Serving Experience
• Have experience deploying a ML models in different deployment architectures; databases, streaming, messaging queues, etc.

The ideal candidate will:
• Have a T-shaped skill set; a strong foundation in ML/AI, computer science, software engineering, statistics, mathematics or related field. Along with a single area of specialization such as NLP, recommendation engines, time-series forecasting, computer vision, etc.
• Building solutions that not merely work, but solutions that permit the fastest deployment, at scale
• Have good intuitions as to what heuristics or data transformations can be applied to data that are motivated by domain expertise
• Through error analysis, identify how to modify the ML training process for the particular task or, whether new data sources and more observations are necessary
• Have experience deploying a ML models in different deployment architectures; databases, streaming, messaging queues, etc.
• Have experience designing data labeling workflows with cloud-based tools, ideally in AWS

PeopleInsight is a collaborative group of product, program, design, science, and engineering experts.

Our Worldwide Operations team numbers in the hundreds of thousands of dedicated associates and leaders. They deliver millions of packages to customers every day. PeopleInsight is the behind-the-scenes team that supports field and organizational leadership by advising action, informing decisions, and identifying risk. The PeopleInsight team resides at the intersection of words and numbers. We listen, observe, and measure at scale to support customer success with insights that are consumable, actionable, and relevant. Our work is dedicated to empowering leaders and enabling action through data and science. In addition to standard reporting, we leverage analytics to help our leaders focus their efforts in ways that will engage, retain and grow their associates to propel the business forward.