Close

Peter Kaiser

Software Engineer

Download Resume

About Me

I am a software engineer with a current interest in Machine learning & Deep learning, I arrived here via a circuitous route.

First I was a talent agent for 3 years, and then an entrepreneur for about 2 years. I am a founder or co-founder of a few companies involved in iOS Mobile development, Web development, VR and AR technologies. Since selling my company in 2015, I have devoted my energies to studying and applying software engineering and machine learning techniques. My enthusiasm for machine learning (data engineering) has made me learn everything from scratch. I have learned from the best top university materials online, and have honed my skills in numerous Kaggle competitions, as well as real-world problems. I strongly believe in continuous learning as well as collaboration.

Experience

ICS

Software Engineer

●Created 50+ proprietary production-ready trading algorithms (growth value, equity momentum, factor value, dual momentum) utilizing Python, Pandas, Numpy and Scikit-learn using company fundamentals and alternative datasets with Alphalens and backtesting software.
●Implemented Decision Tree Classifier, Random Forest, and Recurrent Neural Network NLP for stock sentiment, which resulted in 25% increase in alpha while decreasing volatility by 15%.
●Performed weekly implementation of economic research papers to test new trading algorithms for production.
●Achieved backtested resulted in 28.4% annual return over 10 years, sharpe ratio of 1.10, and Alpha of 18%.


Developed a daily, weekly and monthly trading algorithms, developed news release forecasting algorithms; analyzed algorithms for improvement; backtesting new and existing algorithm changes before implementation.

Algorithms focused on growth style investment strategies. Overseeing investment strategies manages funds and client mandates and active in various private equity ventures. Investments range covers quant hedge funds, real estate and more.

Greenfield Workshop

Software Engineer

Developed and created Virtual Reality experiences from scratch to client specifications

● Built full digital VR museum and other experiences release ready.
● Using photogrammetry, Unity, Maya, Unreal, Aerial drone photography and Airtable database
● Resulted in programmed and modeled experiences over 100,000 square feet of virtual reality. Created International exposure for partnered museums.

Virtual Reality Creative Direction, Experience Design, Prototyping, and Development for non-gaming apps using Maya, Unity, Photoshop, photogrammetry, and other tools to create concept presentations, flows, functional 3d prototypes, and deliverables.

Tangerine Talent

Co-Founder

Talent Agency, representing talent for print, commercial, and theatrical projects

● Led and trained a team of five Talent Agents
● Built client list from 0 to over 300 through an extensive recruitment
● Booked over 100+ project and increase YoY revenue by 125%
● Notable clients: Visa, Apple, Verizon, Samsung, Toyota, Nike, Google, Nokia

I was a Co-Founder and CEO of Tangerine Talent. We went from having less than 10 clients to over 300 clients in the course of 3 years. I worked with Fortune 500 companies and brands to produce over 200 commercials for online media, print, national, and international campaigns. I was in charge of day to day as well as recruiting and client management.

Partial List of Clients and Commercials

Visa, Apple, Verizon, Mcdonalds ,Toyota, Nike, Subaru, Nokia, Ubisoft, Samsung, Google, NFL, NBA, Best Buy, Disney, Logitech, Lincoln, Jordan Nike, 3M, Johnson Baby, Hyundai, Got Milk, One A Day Vitamins, GoGurt, FM Global, Food Network, DropBox, Direct TV, Chevy Van, Bridge Stone, At&t U-Verse, Wolfenstein, Budlight, Budweiser, Converse, SportsCaster, State Farm and Time Warner

Education

University of San Francisco

Deep Learning at Data Institute

This was a Professional Certificate Program in which I learned the practical details of deep learning applications with hands-on model building using PyTorch and fast.ai and worked on problems ranging from computer vision, natural language processing, and recommendation systems.

After finishing this course, I am able to:

apply transfer learning to image classification problems
use neural networks for recommendation algorithms
use recurrent neural networks and convolutional neural networks for text classification problems
apply neural networks to tabular data and learn embeddings for categorical variables

Part 2 tackles more complex problems that require integrating several techniques. This includes both integrating multiple deep learning techniques, as well as combining classic machine learning techniques with deep learning. All methods will be introduced in the context of solving end-to-end real-world modeling problems.

We reimplemented the fast.ai library from scratch, allowing me to have a full understanding of DL.

Hack Reactor

Advanced Immersive Software Engineering Program

The Hack Reactor immersive coding Bootcamp is focused on building autonomous, fully capable software engineers. Eleven hour days, six days a week for 12 weeks focused on full stack Javascript. With 400 hours of lectures and course work and another 400 hours of building full stack real web apps, this program fosters software engineers ready to succeed in today's tech industry.

Udacity

Nanodegree Program - AI for Trading

Quantitative analysis, including data processing, trading signal generation, and portfolio management. Using Python to work with historical stock data, develop trading strategies, and construct a multi-factor model with optimization.

Basics of quantitative analysis, including data processing, trading signal generation, and portfolio management. Use Python to work with historical stock data, develop trading strategies, and construct a multi-factor model with optimization.

Project: Trading with Momentum
Project: Breakout Strategy
Project: Smart Beta and Portfolio Optimization
Project: Multi-factor Model

Learn how to analyze alternative data and use machine learning to generate trading signals. Run a backtest to evaluate and combine top performing signals.

Project: NLP on Financial Statements
Project: Sentiment Analysis with Neural Networks
Project: Combining Signals for Enhanced Alpha
Project: Backtesting

https://confirm.udacity.com/5T9UMFT4

Udemy

Python for Financial Analysis and Algorithmic Trading

Learned the fundamentals of Python, and then proceed to learn about the various core libraries used in the Py-Finance Ecosystem, including jupyter, numpy, pandas, matplotlib, statsmodels, zipline, Quantopian, and much more!

We covered the following topics used by financial professionals:

Python Fundamentals, NumPy for High Speed Numerical Processing, Pandas for Efficient Data Analysis, Matplotlib for Data Visualization, Using pandas-datareader and Quandl for data ingestion, Pandas Time Series Analysis Techniques, Stock Returns Analysis, Cumulative Daily Returns, Volatility and Securities Risk, EWMA (Exponentially Weighted Moving Average), Statsmodels, ETS (Error-Trend-Seasonality), ARIMA (Auto-regressive Integrated Moving Averages), Auto Correlation Plots and Partial Auto Correlation Plots, Sharpe Ratio, Portfolio Allocation Optimization, Efficient Frontier and Markowitz Optimization, Types of Funds, Order Books, Short Selling, Capital Asset Pricing Model, Stock Splits and Dividends, Efficient Market Hypothesis, Algorithmic Trading with Quantopian, Futures Trading

https://www.udemy.com/certificate/UC-GROGCWT6/

General Assembly

User Experience and Swift Programming

Mobile Development

I learned the skills you need to build iPhone apps and submit them to the app store. I developed an understanding of object-oriented programming with Swift, Apple’s new programming language, and learn the best practices for iOS interface design.

User Experience Design

This course was an introduction to User Experience with a focus on Information Architecture. I created a complete UX brief, from ideation to annotated wireframes that could be handed off to a developer to be built. We also learned to conduct the research necessary and create all the supporting documentation required to defend our design decisions. There was a large focus on how to dissect and analyze common design patterns.

Gnomon School of Visual Effects

Particle and Fluid Dynamics

I was in a custom track focused on Dynamic Effects. The track consisted of 24 courses. It started with grounded a traditional background. It then quickly moved into Maya and Houdini. With a focus on particle effects and fluid dynamics.

Projects

Deep Learning Library 

Built library to simplify and speed up training for neural nets using modern best practices 
Built scalable neural net framework with callbacks for customization 
 Implemented LAMB and ADAM optimizer 
 Added custom mixup for CNNs allowing for better training due to variability

Histopathologic Cancer  Detection

Created a model to identify metastatic cancer in small image patches taken from larger digital pathology scans
Built a deep learning model with Pytorch.
Models used densenet201, densenet169, resnet50 and VGG16
Placed 41st top 4% globally with an accuracy of 97.99%

Santander Customer  Transaction Prediction

Created a model to identify which customers will make a specific transaction in the future  
Built a deep learning model with Pytorch and sklearn 
 Used OOF Stratified K-Folds cross-validator and RNN 
 Placed top 14% globally with an accuracy of 89.98%

Skills

Get in Touch