Akhil V

*Built advanced agentic AI systems leveraging multi-agent workflows, LLMs, and memory architectures to support autonomous enterprise decision-making. *Designed and deployed LLM-driven applications using LangChain, CrewAI, and Semantic Kernel to automate intelligent task flows. *Developed Retrieval-Augmented Generation (RAG) pipelines using FAISS and Pinecone for enterprise search and contextual document summarization. *Fine-tuned and evaluated open-source LLMs such as LLaMA, Mistral, and Falcon to improve domain-specific performance and reduce hallucinations. *Applied prompt engineering strategies and chain-of-thought reasoning to optimize LLM output quality and task-specific accuracy. *Integrated GenAI solutions into production with Azure ML, MLflow, Databricks, and Power Platform, including CI/CD pipelines for deployment and monitoring. *Built Copilot agents and custom conversational flows in Microsoft Copilot Studio, integrating with Power Automate and Power Apps for HR, finance, and operations automation. *Collaborated with product and engineering teams to identify and implement GenAI use cases for knowledge retrieval, onboarding, and support automation.

*Developed and deployed machine learning models for classification and regression using Python, Scikit-learn, and XGBoost, improving predictive accuracy by 20%. *Built and fine-tuned NLP pipelines for document categorization and sentiment analysis using spaCy and Hugging Face Transformers. *Designed feature engineering workflows to handle structured and semi-structured data using pandas, NumPy, and SQL. *Automated model training and evaluation pipelines with MLflow, ensuring reproducibility and consistent deployment environments via Docker. *Collaborated with data scientists and engineers to deploy REST API-based models in AWS environments using Flask, EC2, and SageMaker. *Implemented real-time model monitoring via AWS CloudWatch and automated retraining logic based on data drift and performance thresholds. *Contributed to internal tooling for dataset versioning and model explainability using SHAP and LIME. *Communicated model performance through interactive Power BI dashboards for business stakeholders.

*Designed and deployed end-to-end machine learning solutions for tasks including classification, forecasting, recommendation, and anomaly detection using Python, Scikit-learn, TensorFlow, and PyTorch. *Built NLP pipelines leveraging spaCy, NLTK, and Hugging Face Transformers for sentiment analysis, document classification, and named entity recognition (NER). *Engineered and processed large-scale datasets using pandas, NumPy, and SQL to optimize model performance and reduce data preparation time. *Deployed ML models via REST APIs using Flask and FastAPI, containerized with Docker, and orchestrated using Airflow and Kubernetes. *Applied MLOps best practices including CI/CD pipelines, model versioning, and real-time monitoring with MLflow, Prometheus, and AWS CloudWatch. *Created automated retraining pipelines triggered by data drift using AWS Lambda, Step Functions, and S3 versioning. *Built scalable workflows on AWS SageMaker, EC2, and Glue, ensuring cost-effective and performant machine learning operations. *Delivered insights through dynamic dashboards and reports built with Power BI for cross-functional stakeholders.

Currently pursuing a PhD in Artificial Intelligence, focused on advanced coursework and doctoral-level training in AI and machine learning. Building a strong foundation in AI theory, machine learning algorithms, deep learning, and large-scale intelligent systems, while preparing for formal research and dissertation work. Program emphasizes applying AI concepts to real-world, enterprise, and large-scale systems.