I am a Data Scientist with a solid background in developing Python-based data pipelines and creating LLM-driven reporting systems. My experience includes optimizing time-series anomaly detection models using Multivariate Autoencoders and advanced signal processing techniques. I have successfully integrated MCP with Trino for large-scale data processing, effectively addressing challenges like data truncation and pipeline scalability. With a keen eye for detail and a commitment to extracting actionable insights from complex datasets, I bring both creativity and technical acumen to my projects. If you’re looking for someone with hands-on experience in data science who can enhance your data processes and drive impactful results, I would be thrilled to explore your project further.
Add work experience to your profile. (optional)
– Designed and enhanced AI-driven data pipelines for automated report generation, integrating structured (.csv) and unstructured (.eml) data sources with LLMs for analytical insights.
– Conducted data profiling, quality assessment, and cleaning strategies, addressing real-world challenges such as large input truncation and incomplete LLM outputs through prompt engineering and data chunking.
– Researched and applied Model Context Protocol (MCP) and evaluated its integration with Trino to improve large-scale data accessibility, query performance, and system reliability.
– Explored and tested Agentic AI systems and sandbox environments, developing markdown-based workflows for autonomous data processing, reporting, and performance evaluation.
– Implemented workflow automation and orchestration using AI-assisted sandbox tools and Airflow DAG concepts to enable repeatable, traceable data operations.
– Applied time-series analysis and anomaly detection techniques, including denoising strategies and Multivariate Autoencoder (MVA) models to improve sensor data quality and detection accuracy.
– Studied and adapted existing wafer anomaly detection solutions, refining preprocessing, model configurations, and thresholds to support shadow deployment readiness.
Add work education to your profile. (optional)
We will review the reports from both freelancer and employer to give the best decision. It will take 3-5 business days for reviewing after receiving two reports.