Data Scientist

Psybergate is an IT company that builds bespoke software solutions and provides highly skilled resources to its clients.

Our Financial Services client in JHB North seeks a talented Data Scientist to join their team. In this role, you will work closely with Sales, Trading, and IT teams to manage and analyze data from multiple sources to support electronic trading and market analysis. This is a 3-month contract with the possibility of renewal.

Key Responsibilities:

1. Develop and optimize market-making models focused on equities and equity derivatives.
2. Conduct back-testing and research to enhance and refine existing models and algorithms.
3. Analyze high-frequency trading data to identify patterns, trends, and anomalies.
4. Implement machine learning techniques (LSTM, convolutional networks) to predict short-term price movements.
5. Collaborate with cross-functional teams to integrate new indicators and models into production environments.
6. Work with AWS infrastructure (S3, SageMaker) for data processing, model training, and deployment.
7. Generate and analyze data samples, including time-based and volume-based sampling.
8. Ensure the robustness and performance of models by balancing in-sample and out-sample results.

Background and Experience Required:

1. Experience: Minimum 5-7 years in data analytics, machine learning, or quantitative analysis within financial markets.
2. Industry Experience: Proven experience in global markets, trading platforms, financial systems, and applications (e.g., equities, derivatives, trading algorithms).
3. Technical Skills: Strong proficiency in Python, SQL, and experience with AWS (S3, SageMaker).
4. Familiarity with statistical and machine learning techniques, particularly LSTMs and convolutional networks.
5. Experience with high-frequency trading data, market microstructure, and quantitative analysis.
6. Knowledge of trading algorithms and their application in market-making and equities trading.
7. Hands-on experience with back-testing and developing data-driven models.