Michael Manurung Email & Phone Number

Jakarta, Indonesia

• Developed a multi-asset price volatility regression model using gradient boosting, achieving a Mean Absolute Error (MAE) of under 0.5% in volatility prediction. Additionally, leveraged a decoder-only model to forecast future closing prices of assets and their peers.• Created an LLM agent to generate equity research reports by simulating analyst research process. This involved integrating anomaly detection results as part of the feature engineering within the company’s fundamental dataset. The model is capable of providing responses in an average of 15 seconds.• Utilized Redis vector database for similarity search, enabling classification of user queries before processing by an LLM with function calling for similar cases.

Jakarta, Indonesia

• Developed a web-based chat interface for an LLM pilot project, utilizing a React-based framework.• Conducted gap analysis and competitor benchmarking by analyzing data such as site traffic, customer onboarding statistics, and other proxy metrics to visualize industry rankings.• Contributed to the implementation of an LLM-powered landing page targeting Sharia account openings, assisting customers with stock market-related inquiries.

Jakarta, Indonesia

• Developed an ordering feature for bond Initial Public Offerings (IPOs) on the Alternative Market Trading System, while seamlessly integrating it with the firm’s legacy system.• Revamped the firm's trade confirmation system, which handles thousands of transactions and maintains third-party connections through a parallel and event-driven architecture.• Managed terabytes of internal core data backup and ensured secure transfer protocols.• Contributed to the development of the firm's fund locking application, primarily designed for IPO allotment purposes.

Undergraduate, Computer Science

Science