Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

AI Engineer with 3+ years of experience building production Generative AI, RAG, and machine learning systems. Experienced in Python, LangChain, LlamaIndex, FastAPI, AWS, and scalable AI infrastructure, with a passion for developing reliable AI solutions that solve real-world business problems. Enjoys collaborating across teams to turn innovative ideas into production-ready products with measurable impact.…AI Engineer with 3+ years of experience building production Generative AI, RAG, and machine learning systems. Experienced in Python, LangChain, LlamaIndex, FastAPI, AWS, and scalable AI infrastructure, with a passion for developing reliable AI solutions that solve real-world business problems. Enjoys collaborating across teams to turn innovative ideas into production-ready products with measurable impact.

Mahendra Avudiyappan

AI Engineer, Data Scientist, Data Engineer, +2





AI Engineer with 3+ years of experience building production Generative AI, RAG, and machine learning systems. Experienced in Python, LangChain, LlamaIndex, FastAPI, AWS, and scalable AI infrastructure, with a passion for developing reliable AI solutions that solve real-world business problems. Enjoys collaborating across teams to turn innovative ideas into production-ready products with measurable impact.…AI Engineer with 3+ years of experience building production Generative AI, RAG, and machine learning systems. Experienced in Python, LangChain, LlamaIndex, FastAPI, AWS, and scalable AI infrastructure, with a passion for developing reliable AI solutions that solve real-world business problems. Enjoys collaborating across teams to turn innovative ideas into production-ready products with measurable impact.

Available to hire

AI Engineer with 3+ years of experience building production Generative AI, RAG, and machine learning systems. Experienced in Python, LangChain, LlamaIndex, FastAPI, AWS, and scalable AI infrastructure, with a passion for developing reliable AI solutions that solve real-world business problems. Enjoys collaborating across teams to turn innovative ideas into production-ready products with measurable impact.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Work Experience

AI Engineer at AIG

January 1, 2026 - Present

Architected multi-agent decisioning workflows using LangGraph, CrewAI, and FastAPI to parse complex multi-line commercial insurance policy applications, improving underwriting triage velocity. Engineered an enterprise RAG intelligence platform with LlamaIndex, DeepSeek LLMs, and Pinecone vector search over AWS SageMaker, reducing claims resolution cycle times by 34%. Optimized high-throughput production inference pipelines for deep sequence models using PyTorch and Docker, achieving reductions in p95 processing latency while handling millions of transactions via CloudWatch monitoring.

AI Engineer at AIG, United States

January 1, 2026 - Present

Architected multi-agent decisioning workflows using LangGraph, CrewAI, and FastAPI to parse complex multi-line commercial insurance policy applications and accelerate underwriting triage. Built an enterprise RAG platform using LlamaIndex, DeepSeek LLMs, and Pinecone vector search deployed on AWS SageMaker, reducing claims resolution cycle times by 34%. Optimized high-throughput production inference pipelines for deep sequence models using PyTorch and Docker with monitoring via AWS CloudWatch, achieving an executive-mandated reduction in p95 processing latency while handling millions of transactions.

AI/ML Research Assistant at DePaul University

June 1, 2025 - November 1, 2025

Built and deployed a real-time anomaly detection system for pharmacy workflows, reducing delays and improving prescription verification efficiency. Engineered a scalable feature/ETL pipeline with 115K+ records, improving ETL speed by 30% and data reliability. Trained PyTorch deep learning models, improving macro-F1 by 8.9 points and minority-class recall by 15%. Reduced false positives by 22% using class balancing and hyperparameter tuning. Containerized training/inference pipelines with Docker and MLflow, increasing throughput by 11% and reducing queue delays by 18%, and implemented automated model evaluation/validation for production readiness.

AI/ML Research Assistant at DePaul University, Chicago, IL

June 1, 2025 - November 1, 2025

Developed and deployed a real-time anomaly detection system for pharmacy workflows to reduce delays and improve prescription verification efficiency. Engineered 115K+ records into scalable ML feature pipelines, increasing ETL speed by 30% and improving data reliability. Trained deep learning models in PyTorch to improve macro-F1 by 8.9 points and increase minority-class recall by 15%, while reducing false positives by 22% using class balancing and hyperparameter tuning. Containerized training and inference using Docker and MLflow to improve throughput by 11% and reduce queue delays by 18%, and implemented automated model evaluation/validation pipelines for production-grade deployment standards.

Machine Learning Engineer at Informative Web Solutions

October 1, 2021 - November 1, 2023

Constructed fraud and anomaly detection systems using TensorFlow (including autoencoders) and PostgreSQL to proactively screen web traffic and mitigate annualized cyber risk exposure by $1.4M. Built interactive real-time analytics/reporting architectures with Tableau and Power BI, translating forecasting outputs into dashboards that increased stakeholder conversion rates by 19%. Designed distributed feature engineering and ETL pipelines using PySpark, Databricks, and Snowflake, reducing model training time by 40%. Fine-tuned localized open-source language models (BERT, Llama-2) with Hugging Face Transformers and LoRA on GCP Vertex AI for sentiment indexing and improved retention. Implemented A/B testing and statistical hypothesis testing with scikit-learn, LightGBM, and MLflow to validate algorithmic improvements that increased quarterly product sales without increasing acquisition costs.

Machine Learning Engineer at Informative Web Solutions, India

October 1, 2021 - November 1, 2023

Built fraud and anomaly detection systems using TensorFlow, autoencoders, and PostgreSQL to screen incoming web traffic and mitigate annualized cyber risk exposure by $1.4M. Designed real-time business intelligence and performance reporting architectures using Tableau and Power BI, translating forecasting outputs into stakeholder dashboards that increased conversion rates by 19%. Implemented distributed feature engineering and ETL pipelines with PySpark, Databricks, and Snowflake, reducing model training time by 40%. Fine-tuned localized open-source language models (BERT, Llama-2) via Hugging Face Transformers and LoRA on GCP Vertex AI to support sentiment indexing and improve user retention. Designed A/B testing and statistical hypothesis testing pipelines using scikit-learn, LightGBM, and MLflow to validate algorithmic improvements for revenue growth without increasing baseline acquisition costs.