Chaithanya Anne

Analyst Assist

• Designed and deployed a production-grade, enterprise-scalable chatbot that converts natural language queries into SQL and Python for data extraction and chart generation, leveraging GPT-4o.
• Implemented robust guardrails against data breaches and SQL injections, along with in-app context memory, enabling follow-up questions without hallucinations, to maintain conversational continuity.
• Built a secure, Streamlit-based UI with Okta authentication and live user feedback for continuous improvement; it currently supports 60+ massive tables, with 100% charting and 90% SQL accuracy using CoT and RAG techniques.
• Adopted by over 100 users across three business verticals, cutting TAT from two days to two hours for routine analysis, visualization, and business leader requests.

Agentic RCA Workflow

• Built an autonomous, React-based, agentic workflow to identify anomalies in warehouse consumption using Snowflake’s ML-powered Top Insights algorithm, based on automated test/control data generation.
• Engineered dynamic extraction of top contributors from anomalies, triggering agent-driven deep-dive SQL analysis to drill into root causes.
• Integrated LLMs to parse insights and generate RCA summaries, which are auto-emailed to audit teams and affected users based on individual usage attribution.
• Orchestrated end-to-end workflow via Snowflake stored procedures, reducing Fin/Ops audit turnaround from 2 days to 15 minutes.

QE Agent

• Developed an LLM-driven agent to automate SQL test case generation and validation across stored procedures, views, and tables, covering key checks like accuracy, consistency, metadata, and transformations.
• Built a unification layer to deduplicate and generate an incremental regression suite.
• Integrated automated test execution into CI/CD pipelines for seamless pre-production validation of all database changes.
• Reduced manual test case creation by 80%, significantly enhancing QA efficiency and development velocity.

Fin/Ops Optimization

• Integrated AI-based summarization into dashboards through stored procedures for end-user ease of consumption.
• Created an alerting framework for task monitoring on Snowflake, reducing manual effort to zero for observing and diagnosing the errors.

Gen-AI products and POCs

• Secured funding of INR ₹1.2 crore for a 6-month period to accelerate the code-writing chatbot POC from a Fortune 20 retail giant.
• Developed phase 1 of the code-writing chatbot using Google's AI Model Suite, which extracts data and builds dynamic charts in response to natural language questions, reducing TAT for data analysis by 60% in the enterprise sales and marketing team.
• Created a POC on Neo4J knowledge graphs for accurate information retrieval.
• Worked with vector databases and RAG to augment insights, drill through for the Customer Intelligence Team, reducing reporting overhead labor by 75%.
• Created 5+ functional and technical architectures to integrate GenAI into retail data science and analytics workflows.

App Download and Re-engagement Predictive Modelling

• Led a team of five in strategizing the annual roadmap for a Fortune 20 retail company's mobile app growth.
• Created and deployed four propensity models based on XGBoost to identify potential app downloaders and app re-engagers, promoting the newly launched mobile application.
• Achieved a 17% install rate, and a 33% increase in downloads per month, post-campaign launch.

Value Audience for Event Driven Engagement

• Built a sophisticated k-means micro-segmentation clustering model, creating diverse audience cohorts.
• Developed an advanced look-alike modeling system by building XGBoost propensity models on top of the suitable clusters to target high-value audience segments.
• Generated an exceptional ROI, with an average iROAS of $35 across these customer segments.

Retail Media Campaign Strategy and Deployment

• Strategized and deployed over 15 audience campaigns for multiple vendors of a Fortune 20 retail giant, reducing TAT by approximately 57% for RM+ campaign requests.