Ganesh Gowda Venugopal

AI Engineer at TankTideAI

Built document processing AI agents with LangChain integrating AWS Textract for OCR, pgvector for semantic context retrieval, and GPT-4 for draft response generation with human approval gates. Developed a RAG pipeline with pgvector-powered semantic search and LLMs to assist operations and support teams by quickly retrieving relevant resumes, contracts, and ticket history, streamlining knowledge access and reducing manual lookups. Implemented quality gates using Pydantic schema validation, OCR confidence thresholds, and response guardrails to ensure output quality. Orchestrated edge case AI workflows with low code platforms (Make, n8n) to handle exceptions and lightweight integrations outside the core pipeline, reducing engineering overhead.

Lead Squared Machine Learning Engineer at LeadSquared

Deployed a Next Best Action recommender system using causal inference with LightGBM on Databricks and Spark, achieving a 21% increase in lead to customer conversion validated through A/B testing and uplift modeling. Engineered a real-time leads scoring engine by creating behavioral embeddings from CRM interaction sequences (clickstream and communication logs) using PyTorch sequence models and feature engineering. Optimized pipelines with PySpark and Delta Lake. Built reproducible ML pipelines with MLflow for experiment tracking, Docker and Kubernetes for deployment, and CI/CD automation with GitLab. Integrated unit and regression tests for models, reducing deployment failure rate to under 2%. Collaborated cross-functionally with product managers and sales teams to embed ML recommendations into CRM workflows.

Machine Learning Engineer Intern at Startup (CV Processing MVP)

Led development of a CV processing MVP to extract text from PDFs and scanned documents using OCR, built ML data pipelines, and generated structured JSON summaries; implemented tracking with MLflow and created dashboards; collaborated with the data team to ensure reproducibility.