Melissa L.

Leading internal Etsy Ads organization of four product teams focused machine learning, seller experience, and platform that support our promoted search and recommendations products and system. + Supported recruiting, onboarding, and coaching for front line engineering managers and senior staff ICs + Partnering with Product and Analytics Leads to set direction of organization + Lead company wide retrieval infrastructure modernization effort.

Leading organization of three machine learning engineering teams that support our promoted search and recommendations products, including retrieval, embeddings, user engagement models, bidding, pacing, marketplace efficiency, optimization, and ML Ops. + Supported recruiting, onboarding, and coaching for front line engineering managers and senior staff ICs + Responsible for an organization of 35 + Partnering with Product and Analytics to set direction of organization + Led efforts to redesign hiring for ML talent across company Languages/Tools: Python, Scala, Tensorflow, pytorch, Airflow, Docker, GCP

Leading two machine learning engineering teams that support our promoted search and recommendations products, including retrieval, embeddings, joint optimization, and user engagement models. + Grew Ads Quality Retrieval and Ranking efforts from a single team to multiple teams with deeper product focus for our promoted search and promoted recommendation products. Languages/Tools: Python, Scala, Tensorflow, pytorch, Airflow, Docker, GCP

Leading machine learning engineering team that support our ads marketplace with a focus on user engagement models, bidding, pacing, marketplace efficiency, and ML Ops. + Built out Ads Marketplace team, defined technical and strategic roadmap, grew team by 50% and led hiring efforts. Languages/Tools: Python, Scala, Tensorflow, pytorch, Airflow, Docker, GCP

Delivery Tech Lead for Fraud Machine Learning Application + Led the full lifecycle of Third Party Fraud ML Solution, preventing $15M in losses, including technical roadmap planning with business and product partners, developing ML models, managing and orchestrating production ETL pipelines; deployment, monitoring and support of model scoring applications, and model governance. + Led two teams of engineers, split across model development and model deployment. Languages/Tools: Python, Spark, H2o, Flask, Kafka, Jenkins, Docker, Kubernetes, gRPC, Airflow, AWS (EMR, Lambda, S3)

Technical Lead for Third Party Fraud ML Applications + Led team of four engineers in development and deployment of ML applications. + Built cloud-based production ETL pipelines and model scoring applications for streaming and batch models to support multiple fraud initiatives, including graph based approaches. + Partnered with business and data science partners to assess production viability and managed cross team dependences with partner tech teams to ensure proper execution. Languages/Tools: Python, Java, Spark, Docker, Kubernetes, gRPC, AWS (EMR), Airflow

Machine Learning and Software Engineer in Capital One’s Center for Machine Learning, focused on multiple consulting and research engagements. Primary focus on fraud risk modeling. + Developed POC Time Series Neural Network Based Model + Researched Explainability for AL/ML, resulting in publications at NeurIPS 2018 FEAP and ACM/AAAI AIES + Led ETL pipeline creation for First Party Fraud risk model. Managed ETL orchestration that served for model development and model deployment. Developed model packaging and deployment strategy. Partnered in the development of DevOps CI/CD pipelines across ETL and Model Serving Applications. Languages/Tools: Python, Spark, TensorFlow, H2o, sklearn, AWS (EMR, Step Functions, Lambda), Jenkins

Developed BKrawl.nyc web app that allows users to create custom bar crawls in Brooklyn. BKrawl utilizes OpenStreetMap API to collect street and intersection data and the Untappd API to locate bars. Machine learning techniques are used to determine localized regions with high crime rates using kernel density estimation and shortest path with the Dijkstra algorithm. MySQL and Python back-end to find shortest path between bars while avoiding scenes with high traffic incident rates and crimes. Designed and coded interactive front-end using Flask, Bootstrap, and Leaflet, deployed on AWS. Project completed within three weeks.

For my dissertation research, I studied extragalactic star formation in nearby spiral galaxies. Duties and Highlights: Analyzed and connected multi-wavelength data sets. Lead the primary reduction and calibrator of over 600 hours of data for international collaboration. Principle Investigator of two separate proposals, one including funding. Published two first author papers.

Taught two lab sections for a full year of Introductory Physics at Stony Brook University. My duties included preparing a short lecture to review the topics covered in the lab, assisted students in the lab and verbally evaluated students performance in the lab at the end of the class. In addition, I continued to present guest lectures and special topics seminars for various astronomy classes.

Operated a 14 inch telescope at Drew University. Duties included using the telescope for weekly public viewing nights, guests included students of the university, school groups, and members of the local community. The position also included working as a teaching assistant for the Introductory Astronomy courses offered once a year.

Mentor: Dr. James Miller-Jones Project: Searching for Jet-ISM Interactions Around X-ray Binary Systems. Data: Archival VLA data

Mentor: Dr. Todd Hillwig Project: Photometry of Binary Central Stars of Planetary Nebulae Data: Optical observations taken with the SARA .9m telescope on Kitt Peak