Hire Rahul R. - Fine-Tuning Engineer

* Thesis: Towards Small Language Models for Security Query Generation in SOC Workflows * Designed a two-stage SLM-LLM architecture for Kusto Query Language (KQL) generation, reducing estimated token costs by 15x compared to state-of-the-art methods while achieving 97% syntactic correctness. * Performed Supervised and Chain-of-Thought (CoT) Fine-Tuning of SLMs using Performance-Efficient Fine Tuning (PEFT) and Low-Rank Adaptation (LoRA), achieving KQL syntactic correctness up to 85%. * Replicated and benchmarked LLM-based data engineering pipelines to quantify the effectiveness of published, state-of-the-art methodologies. * Developed evaluation metrics to assess the semantic validity and accuracy of LLM-generated KQL queries.

* Developed user-friendly Dash applications, Kibana, and Tableau dashboards to effectively communicate geospatial analytic insights. * Created Data Engineering pipelines using Python and Bash scripts to assist technical analysts with processing large datasets, significantly reducing manual review of information. * Refactored existing Extract, Transform, Load (ETL) codebases to promote code reusability and efficiency, decreasing program execution time. * Mentored new employees by performing regular code reviews, and answering questions related to Time Series Analysis/Autoregressive Integrated Moving Average (ARIMA) modeling, Natural Language Processing (NLP), Geospatial Data Engineering, and Social Network Analysis.