Software Engineer | Machine Learning
2022 - PRESENTGoogle- Performed relevant work in a horizontal machine-learning team.
- Collaborated with various stakeholders in different organizations from my own.
- Built software which resulted essential to the company.
Technologies: Artificial Intelligence (AI), Python, C++, SQL, HyperparametersSenior Machine Learning Engineer
2020 - PRESENTKUNGFU.AI Advanced Data Science Services- Characterized behavior of an unsupervised learning model meant for government use and implemented code and tests for its deployment.
- Developed project plan and tracked using Asana with agile principles.
- Delivered regular status updates to clients and facilitated requirements discussions.
Technologies: Google Cloud Platform (GCP), TensorFlow, PyTorch, Python 3Deep Learning Engineer (Hybrid Data Scientist and Data Engineer)
2018 - 2020Node.io- Gave Node.io its first real AI capabilities--built a predictive modeling and analytics stack from the ground up with scalable AWS (ECS, EC2, Docker) and ML (feature engineering, RNNs, DNNs, and classifiers using PyTorch, XGBoost, Pandas) tools.
- Improved Node's capabilities with new models and new features (e.g., LSTM/GRU sequence models, denoising autoencoder models for data enhancement and neural embeddings).
- Consulted on coworker projects and acted as resident research-paper-reader.
Technologies: PyTorch, Linux, Amazon Web Services (AWS), Amazon, PythonResearch Assistant/Research Engineer
2012 - 2018Oracle Labs- Expanded the market for Oracle's big data analytics offering by providing cutting-edge fraud detection capabilities.
- Expanded Oracle PGX's analytics market share by building a linearly-scalable asynchronous query engine for its distributed execution mode (see GRADES17 paper).
- Introduced Oracle PGX graph analytics to the big data market by building its first large-scale distributed execution mode in C++ and Node.js.
- Gave fine-grained performance optimization capabilities to the Oracle Database team by writing a custom Linux kernel.
- Provided the option to run Oracle Database in a safe Java-like environment with on-demand profile-guided optimization using a dynamic C/C++ LLVM runtime.
- Demonstrated and designed a new research product for selling big-data tools; a graphical system for creating dataflow graphs.
- Demonstrated Oracle Coherence to customers for writing message-oriented middleware by building a highly scalable Java application server.
- Showcased Oracle Labs' new Truffle/GraalVM compiler technology for optimized, scalable big data operations in legacy languages by writing an HTML5 programmer's notebook in Node.js.
Technologies: Oracle RDBMS, Linux Kernel, C, C++, Linux, TensorFlow
Nicholas Roth
Big Data Developer in Austin, TX, United States
Member since March 10, 2020
Nicholas is an efficient machine learning engineer and data scientist with experience in data curation, classical machine learning algorithms, and statistical modeling. Starting an engagement, he searches for data and works with the customer on what to build. Once an algorithm and intuition show the most crucial areas, Nicholas shifts gears to get the "juice" out of that data before testing and deploying models in a client's infrastructure, delivering exceptional and detail-oriented work.
Nicholas is now available for hire
Portfolio
- GoogleArtificial Intelligence (AI), Python, C++, SQL, Hyperparameters
- KUNGFU.AI Advanced Data Science ServicesGoogle Cloud Platform (GCP), TensorFlow, PyTorch, Python 3
- Node.ioPyTorch, Linux, Amazon Web Services (AWS), Amazon, Python
Experience
- Linux 9 years
- Big Data 7 years
- Python 6 years
- Algorithms 5 years
- Artificial Intelligence (AI) 5 years
- Machine Learning 5 years
- Natural Language Processing (NLP) 3 years
- Data Science 3 years
Availability
Part-time
Preferred Environment
Vim Text Editor, Jupyter, Linux, Amazon Web Services (AWS), Pandas, NumPy, PyTorch, TensorFlow
The most amazing...
...project I've designed and built is a system for the startup Node.io that trains deep learning models for non-technical customers.
Employment
Experience
- Embeddings for Musichttp://ambii.io
I worked with a team to scrape music video playlists from YouTube and used the data to learn Doc2vec embeddings with my custom implementation of the algorithm in TensorFlow. To build a music recommender system from this, I took an existing “playlist vector” from a user’s listening history and sorted the songs in our dataset by cosine distance, returning the top ones. Before leaving, I suggested that the next step might be to use content-based embeddings derived from the hidden layer of an LSTM trained on the videos’ soundtracks.
Skills
Languages
C++, Python, C, Python 3, SQLLibraries/APIs
PyTorch, NumPy, Pandas, TensorFlowOther
Machine Learning, Artificial Intelligence (AI), Big Data, Algorithms, Natural Language Processing (NLP), Linux Kernel, HyperparametersParadigms
Data Science, AutomationPlatforms
Linux, Google Cloud Platform (GCP), Amazon, Amazon Web Services (AWS)Tools
Jupyter, Vim Text EditorStorage
Oracle RDBMS
