Kostas Nikolakopoulos
Quant Developer
Kostas is a data scientist and quantitative analytics specialist focusing on developing predictive models using machine learning techniques. He worked with multiple clients in the financial services sector in projects such as future balance predictions, credit risk modeling, and simulation engines. Kostas has extensive coding experience in Python, R, and C++ and academic background in theoretical physics with a doctoral degree from Sussex University and an MSc degree from Imperial College.
Portfolio
Availability
Preferred Environment
TensorFlow, Jupyter Notebook, RStudio, PyCharm, Visual Studio
The most amazing...
...performance increase I've implemented brought down the time needed to run a model from days to minutes.
Work Experience
ML Engineer
PepsiCo Global - Main
- Developed the Bayesian model for ad allocation using a very long list of features from ad campaigns.
- Trained and maintained on AWS of the Bayesian model for ad allocation.
- Performed analysis and presentation of the outcomes of the model, specifically feature importance, p-values, etc.
Software and Data Engineer
Reddit, Inc.
- Developed algorithms to optimize advertisement revenue.
- Implemented model changes in Scala to include new features.
- Researched increasing revenue through better budget pacing techniques.
Data Scientist
Grata Inc
- Created an MVP for estimating a company's value from public data.
- Participated in data gathering and building data pipelines.
- Delivered a stand-alone Python app to run an ML model for live valuations.
Python Quantitative Researcher
Tickup (Algo Fund)
- Designed and developed an algorithmic trading platform that connected multiple development components and technologies.
- Onboarded trading strategies from quant workspaces into the trading platform. Performed backtesting and parameter optimization under different scenarios.
- Cleaned, managed, and consolidated data prepared for the go-live version.
Credit Risk Quant
Bank of America Merrill Lynch
- Delivered an IRC/CRM regulatory project dictated by Brexit migration requirements.
- Enhanced aspects of the model to better reflect theoretical requirements and historical behavior. Conducted statistical tests and submitted them to the validation department.
- Improved the performance of the model implementation. Identified the current model's properties, which reduced the execution from days to hours.
Flow Rates Quant
BNP Paribas
- Contributed to the pricing and risk platform of an electronic transformation project.
- Implemented prices and risk-across-rates products such as swaps, bonds, and futures.
- Enhanced the C++ library for pricing and risk calculations.
Quant Developer
Bank of America Merrill Lynch
- Collaborated with the model performance team to backtest the bank models for all asset classes.
- Improved and enhanced the Python codebase and user interface.
- Used Python and C++ coding to simulate risk factors and correlations, applied for calculating profit and loss, XVA, and margins.
Behavioral Modeler
Royal Bank of Scotland
- Led the behavioral modeling team in preparation for separating the Williams & Glyn division of the Royal Bank of Scotland.
- Developed predictive behavioral models for residential mortgages and current or savings accounts. The models' owner was the treasury, using them for the purposes of funds transfer pricing (FTP) and interest rate risk management.
- Coordinated the development of the Python library for the team and developed a web-based GUI for business users to run the models.
Python Quant Modeler
Barclays Bank, PLC
- Developed predictive behavioral models for various portfolios of the bank's investment, corporate, and retail parts using historical time series data; was personally responsible for the residential mortgage book and the corporate term loans book.
- Managed the full lifecycle of the models, from data cleaning to presentation and documentation of results. Performed ad-hoc statistical analyses, scenario analyses, backtesting, and model reviews.
- Contributed to the quant analytics grad training, gaining exposure to all bank departments.
- Performed ad-hoc statistical modeling and statistical data analysis for various projects of the team.
Consultant
d-fine, Ltd.
- Conducted current accounts modeling for a major bank based in Vienna.
- Developed a supervisory mechanism for the EU bank regulator.
- Gained exposure and experience in the large-scale application architecture.
Experience
House Price Prediction
The client can input their property characteristics such as postcode, number of bedrooms, or garden, and get the estimated price. There is also an add-on feature for trend predictions based on property characteristics.
Sports Arbitrage App
Skills
Languages
Python, C++, R, SQL, Java, Go, Python 3, Scala, C#, Perl
Libraries/APIs
TensorFlow, Scikit-learn, Pandas, PyTorch
Tools
PyCharm, Visual Studio, BigQuery
Paradigms
Data Science, Quantitative Research, ETL
Platforms
RStudio, Jupyter Notebook, Amazon Web Services (AWS), Docker
Other
Mathematics, Mathematical Modeling, Statistical Modeling, Statistical Data Analysis, Data Analytics, Statistics, Data Modeling, Machine Learning, Artificial Intelligence (AI), Quantitative Analysis, Simulations, Web Scraping, Bayesian Inference & Modeling, APIs, Data Engineering, Quantitative Modeling, Quantitative Finance, Finance, Distributed Systems, Software Engineering, Numerical Analysis, Algorithms, Back-end Development, Google BigQuery, Machine Learning Operations (MLOps), Deep Learning, Bayesian Statistics, Economics, Natural Language Processing (NLP), Financial Forecasting, GPT, Generative Pre-trained Transformers (GPT), Data Scientist, OCR
Frameworks
Spark
Storage
Data Pipelines
Education
Doctoral Degree in Theoretical Physics
University of Sussex - Sussex, UK
Master's Degree in Theoretical Physics
Imperial College London - London, UK
Bachelor's Degree in Physics
Aristotle University of Thessaloniki - Thessaloniki, Greece