Atharva Vichare

Built automated data pipelines using Snowflake and Python to prepare datasets for LLM fine-tuning. Designed RAG workflows using LangChain with OpenAI APIs and vector search, supporting faster document retrieval and AI-assisted claims processing. Implemented anomaly detection and automated error-log classification with monitoring to reduce manual incident handling. Improved Snowflake performance, added Git/Jenkins CI/CD, and created SLA monitoring and GPT-based compliance summarization.

Architected data lakehouse solutions combining data lake and warehouse patterns using Snowflake and Amazon Redshift. Built large-scale ETL pipelines with Talend and Informatica, including PySpark/Python processing for semi-structured and log data. Implemented automated orchestration and ML workflow tracking with Apache Airflow, and added Kafka-based streaming ingestion for real-time analytics. Delivered CI/CD for ETL/ML releases with Git/Jenkins, created data-quality/drift validation frameworks, and supported production troubleshooting and monitoring dashboards.

Developed a customer-facing portal using Node.js and JavaScript with AWS services such as EC2 and S3. Implemented RESTful backend APIs for billing, payments, and analytics, integrating with BI dashboards for near real-time customer insights. Set up automated functional and regression testing using Jest and Selenium and designed Jenkins-based CI/CD pipelines for frequent releases. Optimized API performance, supported AWS migration of legacy components, and designed database schemas using MySQL and PostgreSQL for microservices.

Developed an auto-analytics platform with a natural-language-to-SQL interface powered by GPT-4 for self-service reporting. Created AutoML-style pipelines and automated dataset summaries and visualization suggestions for large data volumes. Added real-time anomaly detection using LSTM/Isolation Forest approaches to support operational monitoring.

Built a RAG-based IT support ticket resolution system to shorten time-to-resolution using semantic document retrieval. Indexed a large document set for search and improved technical answer quality through prompt engineering and retrieval tuning. Packaged and deployed the application using Docker and a FastAPI service to run the inference pipeline efficiently.