Kev in

Built an LLM twin agent for customer support to improve conversion rate by 23% using Python, PySpark, OpenAI, Selenium, LangChain, and Kafka. Evaluated and monitored candidate pipelines, managing trade-offs with Databricks and AWS. Designed and developed high-performance backends for gaming platforms using FastAPI and Django, including scalable REST APIs for real-time multiplayer experiences with caching (e.g., Redis) to reduce latency. Engineered multiple chatbot projects using Python, FastAPI, NLP, advanced RAG techniques, vector databases, and graph databases. Integrated Cloud Functions and Lambda for serverless event-driven applications processing millions of daily requests.

Built an LLM twin agent to enhance customer support and improve conversion rate by 23%, using Python, PySpark, OpenAI, Selenium, LangChain, and Kafka. Evaluated and monitored candidate pipelines with benchmarking, managing trade-offs with DataBricks and AWS. Designed high-performance backends for gaming platforms using FastAPI and Django; implemented scalable REST APIs for real-time multiplayer gaming and integrated caching (e.g., Redis) to reduce latency. Engineered multiple chatbot projects with advanced RAG, vector database and graph database approaches, and integrated serverless components using Cloud Functions and AWS Lambda to process millions of daily requests.

Integrated Apache Airflow with AWS to monitor multi-stage ML workflows on Amazon SageMaker and used S3 bucket and QuickSight for data querying and visualization. Reduced data retrieval times using Google Cloud Storage and Google BigQuery for large datasets used in internal analytics. Created Dockerfiles for multiple endpoints, validated via Docker-centr al runner (Docker central tooling) with Jenkins jobs, and deployed the jobs to Kubernetes. Developed and deployed a REST web service using Python and Django, achieving a 25% reduction in server response time and a 15% increase in user satisfaction.