Lokesh Soni

Designed and developed AI chatbots using RASA Framework, implementing Semantic Search techniques integrated with Azure OpenAI, Cognitive Search, and Qdrant vector database. Developed an RNN-LSTM model for Autotext completion. Managed the end-to-end lifecycle of ML models and APIs, including development, training, testing, and deployment. Developed ETL data pipelines and infrastructure utilizing Apache Airflow.

Built a comprehensive Demand Forecast initiative using advanced Hierarchical time series analysis techniques (Holt-Winters, ETS, ARIMAX) to predict sales. Implemented SARIMAX modeling to predict influenza cases. Engaged with stakeholders in European, Nordics, and Middle Eastern regions, leading ML-based projects for clients.

Applied Supervised ML techniques (Random Forest, XGBoost, SVM) to segment healthcare providers. Conducted Data Clustering analysis using K-means, K-prototype, Hierarchical, and DBSCAN. Automated daily/monthly ETL jobs using Python and Bash scripts, and generated reports and dashboards to monitor KPIs.