Large Language Models Engineer
Demonstrated experience with Python.
Demonstrated experience with Java.
Demonstrated experience with R.
Demonstrated experience with SQL.
Demonstrated experience with C/C++.
Professional profile
Developed LLM-powered agents tailored to reason, plan, and use tools to automate knowledge retrieval and decision support, and built LLM-as-a-judge evaluation workflows to assess agent performance in the absence of ground truth.. Developed an end-to-end raw material price forecasting system using univariate and multivariate time-series models using frameworks such as skforecast, and built an OCR + LLM pipeline to extract sentiment signals from industry reports and incorporated them as exogenous variables in forecasting models.
Developed a time-series-based malware detection and anomaly-detection system for cybersecurity applications, using PyTorch, sktime, and tsai to analyze hardware performance counter (HPC) data.. Built a sandboxed Linux container for malware monitoring and ETL pipeline to preprocess and transform HPC data for real-time machine learning analysis.. Applied Stackelberg Security Games, enabling optimized decision-making for malware detection and response.
Developed a detection system using time-series based ML algorithms to classify network traffic and identify DDoS attacks.. Implemented models such as ARIMA and LSTM for accurate, real-time anomaly detection and validated system performance with real-world network traffic data and improved the training efficiency by 10%.
Analyzed and visualized the 2020 university entrance exam data for pattern recognition and clustering.. Used unsupervised learning approaches to identify trends in student performance and achieved close to 90% accuracy in predicting national exam scores using TensorFlow and Keras for model training.