I help companies build machine learning systems that are scalable and dependable. I have worked in food delivery, bot mitigation, military systems and have a PhD in electrical engineering with a focus on signal processing
WORK EXPERIENCE
Senior Machine Learning Engineer; Grubhub. Remote (Dec. 2018 - present)
-
Member of Order Volume Forecasting team. Owned and maintained software to predict weekly order volumes for all Grubhub regions.
-
Implemented parallel training and forecasting for models reducing training time by half and allowing parallel forecasting on all regions.
Machine Learning Technical Lead; Distil Networks. Remote (Sept. 2017 – Nov. 2018)
-
Design and build a "serverless" machine learning system to classify biometric data gathered from web clients - features like mouse movement and keyboard interactions are streamed into the classifier (several hundred features/second), classified, stored, aggregated and served to several hundred endpoints at 15 requests/sec all with rapidly scalable serverless infrastructure (AWS Lambda, Kinesis, DynamoDB and API Gateway).
-
Working with team to build data validation around key fingerprinting technology and move the logic into a pluggable module that allow for back-processing of historic data using Spark.
Data Science Team Lead; Distil Networks. Remote (Jan. 2016 – Aug. 2017)
-
Lead project to replace our AWS EMR-based classification system with a Storm cluster. Organized team of 5 people from Data Engineer, Operations and Data Science to create cluster, write Storm topologies and Kafka endpoints for a real-time streaming classification system of client behavior across our global network. System classifies 4k client records/second to identify malicious bots.
-
Managed team of 3 Data Scientists with weekly 1-on-1’s, tasking, setting goals and executing major projects.
-
Team worked on 28 research initiatives to provide the business with areas for bot detection product improvements and methods for finding attackers.
Data Scientist; Distil Networks. Remote (May. 2014 – Dec. 2015)
-
Researched and prototyped Distil's first Machine Learning product for automated detection of bot-like activities on our global network. Designed and tested the model and built the map-reduce-based system to continuously classify incoming records from our global network and deliver to edge network systems.
-
Worked with marketing to collect data for our annual “Bot Report”. Helped to frame the questions they wanted to answer and aggregate the data from billions of records. Wrote Hive/Impala queries to collect data from S3 via an EMR cluster and gathered into spreadsheets so that Marketing to could understand and use the data.
-
Wrote a bayesian network model to determine network interactions to help troubleshoot a data center issue. Collected statistics on machine CPU, Memory, and Network usage and wrote a bayesian network model using the PC algorithm to help determine the causal interactions of the devices.
-
Wrote a process monitoring system to determine when our classification EMR cluster went down using AWS SQS to receive heartbeat messages from the live systems and stored in Redis with TTLs to determine when a system went down.
Systems and Algorithms Engineer; 3 Phoenix, Inc. Wake Forest (May 2012 – April 2014)
-
Developed software in MATLAB for a machine learning algorithm to classify EW/radar signals in littoral environments using a kernel-based elastic network. Presented results to government customer.
-
Developed, tested and deployed algorithm for fusing sonar and radar tracks to remove surface/sub-surface ambiguity of targets. Tracks are fused based on Bhattacharyya distance of PDFs. System eliminated false positives for tracking algorithm and allowed greater sensitivity to true weapon shots.
-
Researched and wrote two Phase 1 Small Business Innovative Research (SBIR) proposals. Was involved in planning and editing of 8 Phase 1 SBIR topics involving tracking, spectral estimation, image processing and deep learning.
PhD Student; Dr. John Muth, NC State University (2006 – 2012)
-
Wrote numerical ray-tracing simulation of the underwater light field for communication applications. Processed millions of photons/min with MATLAB parallel software that ran on Amazon EC2 cloud.
-
Real-time streaming video link implemented with GNU Radio running in Linux. Utilized LED, modulating retroreflector, or diode laser link for transmitter. Transmitted up to 4 Mbps with GMSK modulation.
PUBLICATIONS/PRESENTATIONS:
-
Conference Presentation: “Serverless Machine Learning in Production”. PyGotham 2018.
-
Conference Presentation: “Intuitive Understanding of the Fourier Transform and FFTs”. OSCON 2014 (170,000+ YouTube views)
-
Conference Presentation: “Data Science Behind Bot Blocking”. Data Science Summit 2015
-
Journal Article: “Simulating Channel Losses in an Underwater Optical Communication System”. Apr 2014 Journal Optical Society of America A
-
Dissertation: “Simulation, Modeling, and Design of Underwater Optical Communication Systems”. NC State University
-
Conference Paper: “Underwater optical communication using software defined radio over led and laser based links”. MILCOM 2011
-
Conference Paper: “A MEMS Blue / Green Retroreflecting Modulator for Underwater Optical Communications”. OCEANS 2010
SKILLS
-
Software Engineering - Python, Matlab, Git, software testing, Linux environment, Docker, Kubernetes, AWS Services, Serverless design
-
Data Science - Supervised and unsupervised ML, Scikit-Learn, Spark
-
Data Engineering - SQL, Kafka, Kinesis, Spark, DynamoDB, Redis, Airflow
-
Interpersonal - extremely comfortable with public speaking, enjoys writing, works well in cross-functional teams, can present and communicate to diverse audiences
EDUCATION
North Carolina State University, Raleigh, NC
-
Ph.D. in Electrical Engineering (2/2012)
-
M.S. in Electrical Engineering (12/2007)
-
B.S. in Electrical Engineering (06/2006)
I help companies build machine learning systems that are scalable and dependable. I have worked in food delivery, bot mitigation, military systems and have a PhD in electrical engineering with a focus on signal processing
WORK EXPERIENCE
Senior Machine Learning Engineer; Grubhub. Remote (Dec. 2018 - present)
- Member of Order Volume Forecasting team. Owned and maintained software to predict weekly order volumes for all Grubhub regions.
- Implemented parallel training and forecasting for models reducing training time by half and allowing parallel forecasting on all regions.
Machine Learning Technical Lead; Distil Networks. Remote (Sept. 2017 – Nov. 2018)
- Design and build a "serverless" machine learning system to classify biometric data gathered from web clients - features like mouse movement and keyboard interactions are streamed into the classifier (several hundred features/second), classified, stored, aggregated and served to several hundred endpoints at 15 requests/sec all with rapidly scalable serverless infrastructure (AWS Lambda, Kinesis, DynamoDB and API Gateway).
- Working with team to build data validation around key fingerprinting technology and move the logic into a pluggable module that allow for back-processing of historic data using Spark.
Data Science Team Lead; Distil Networks. Remote (Jan. 2016 – Aug. 2017)
- Lead project to replace our AWS EMR-based classification system with a Storm cluster. Organized team of 5 people from Data Engineer, Operations and Data Science to create cluster, write Storm topologies and Kafka endpoints for a real-time streaming classification system of client behavior across our global network. System classifies 4k client records/second to identify malicious bots.
- Managed team of 3 Data Scientists with weekly 1-on-1’s, tasking, setting goals and executing major projects.
- Team worked on 28 research initiatives to provide the business with areas for bot detection product improvements and methods for finding attackers.
Data Scientist; Distil Networks. Remote (May. 2014 – Dec. 2015)
- Researched and prototyped Distil's first Machine Learning product for automated detection of bot-like activities on our global network. Designed and tested the model and built the map-reduce-based system to continuously classify incoming records from our global network and deliver to edge network systems.
- Worked with marketing to collect data for our annual “Bot Report”. Helped to frame the questions they wanted to answer and aggregate the data from billions of records. Wrote Hive/Impala queries to collect data from S3 via an EMR cluster and gathered into spreadsheets so that Marketing to could understand and use the data.
- Wrote a bayesian network model to determine network interactions to help troubleshoot a data center issue. Collected statistics on machine CPU, Memory, and Network usage and wrote a bayesian network model using the PC algorithm to help determine the causal interactions of the devices.
- Wrote a process monitoring system to determine when our classification EMR cluster went down using AWS SQS to receive heartbeat messages from the live systems and stored in Redis with TTLs to determine when a system went down.
Systems and Algorithms Engineer; 3 Phoenix, Inc. Wake Forest (May 2012 – April 2014)
- Developed software in MATLAB for a machine learning algorithm to classify EW/radar signals in littoral environments using a kernel-based elastic network. Presented results to government customer.
- Developed, tested and deployed algorithm for fusing sonar and radar tracks to remove surface/sub-surface ambiguity of targets. Tracks are fused based on Bhattacharyya distance of PDFs. System eliminated false positives for tracking algorithm and allowed greater sensitivity to true weapon shots.
- Researched and wrote two Phase 1 Small Business Innovative Research (SBIR) proposals. Was involved in planning and editing of 8 Phase 1 SBIR topics involving tracking, spectral estimation, image processing and deep learning.
PhD Student; Dr. John Muth, NC State University (2006 – 2012)
- Wrote numerical ray-tracing simulation of the underwater light field for communication applications. Processed millions of photons/min with MATLAB parallel software that ran on Amazon EC2 cloud.
- Real-time streaming video link implemented with GNU Radio running in Linux. Utilized LED, modulating retroreflector, or diode laser link for transmitter. Transmitted up to 4 Mbps with GMSK modulation.
PUBLICATIONS/PRESENTATIONS:
- Conference Presentation: “Serverless Machine Learning in Production”. PyGotham 2018.
- Conference Presentation: “Intuitive Understanding of the Fourier Transform and FFTs”. OSCON 2014 (170,000+ YouTube views)
- Conference Presentation: “Data Science Behind Bot Blocking”. Data Science Summit 2015
- Journal Article: “Simulating Channel Losses in an Underwater Optical Communication System”. Apr 2014 Journal Optical Society of America A
- Dissertation: “Simulation, Modeling, and Design of Underwater Optical Communication Systems”. NC State University
- Conference Paper: “Underwater optical communication using software defined radio over led and laser based links”. MILCOM 2011
- Conference Paper: “A MEMS Blue / Green Retroreflecting Modulator for Underwater Optical Communications”. OCEANS 2010
SKILLS
- Software Engineering - Python, Matlab, Git, software testing, Linux environment, Docker, Kubernetes, AWS Services, Serverless design
- Data Science - Supervised and unsupervised ML, Scikit-Learn, Spark
- Data Engineering - SQL, Kafka, Kinesis, Spark, DynamoDB, Redis, Airflow
- Interpersonal - extremely comfortable with public speaking, enjoys writing, works well in cross-functional teams, can present and communicate to diverse audiences
EDUCATION
North Carolina State University, Raleigh, NC
- Ph.D. in Electrical Engineering (2/2012)
- M.S. in Electrical Engineering (12/2007)
- B.S. in Electrical Engineering (06/2006)
You have done such a great job! The create infographic resume is best online site where you can get all tips of writing resume.
ReplyDelete