Ayush Trivedi

About Ayush Trivedi

I build ML and GenAI systems that automate safety audits, slashing review times by 70–95%, eliminating thousands of manual hours yearly, and reducing site injuries across North America by streamlining compliance workflows.Key Systems and ResultsAWS Textract + custom OCR extracts document signals, cutting data entry errors by automating manual reviews.Hugging Face Transformers classifies cases accurately, enabling faster, predictable triage over human variability.Python rule engines validate fields; Lambda/Step Functions trigger actions—delivering first-pass yield improvements and average handle time reductions.Approach and ScaleProduct management meets ML: I map workflows, target pain points, build modular pipelines (ingestion → AI interpretation → actioning), and measure latency, accuracy, and human-touch reduction to handle hundreds of weekly cases reliably.​Real-World DeploymentsPython micro-frameworks replace Excel processes; GenAI engines handle multi-document validation; RAG/LLM layers clarify safety classifications—turning operations into SaaS-like tools that boost insights, shorten cycles, and prioritize high-value decisions over repetitive tasks.

Experience

1. Program Management

   Amazon

   Dec 2023 — Present · Bengaluru, IN

   Led the design and deployment of data pipelines supporting ML model development for customer behavior analysis and transaction pattern mining, aligned with product roadmaps in e-commerce and fintech environments.Drove the development and validation of regression and classification models to inform key product decisions such as dynamic pricing, personalized recommendations, and fraud prevention—collaborating across product, data science, and engineering teams.Established MLOps practices by implementing model lifecycle automation using MLflow, Git, and Docker—streamlining release cycles and ensuring scalable ML delivery aligned with engineering sprints.Transformed business objectives into ML solutions by collaborating with cross-functional stakeholders, delivering measurable improvements in personalization, engagement, and retention metrics.Shaped customer experience strategy through the development and refinement of NLP models using Hugging Face Transformers, enabling accurate sentiment analysis and actionable insights from customer feedback.Utilized AWS services (SageMaker, Lambda, S3) to architect real-time ML pipelines, integrating seamlessly with product features and driving faster decision-making at scale.Supported A/B testing and model performance monitoring, using experimentation frameworks to evaluate user impact and ensure feature reliability in production environments.

Education

• Dr. A.P.J. Abdul Kalam Technical University

  Bachelor of Technology (B.Tech.), Electronics and Communications Engineering

  2016 — 2020