Hire Manish A. - Generative AI Architect

* Led the transformation of AI and General AI business needs into practical technology applications, focusing on cutting-edge Enterprise Large Language Models (LLMs) such as Gemini1.5, OpenAI, and Antropic Claudi AI. * Designed strong technical frameworks that integrate various AI models and technologies, including Vector DB for embeddings and Azure Open AI for LLM summarization modules. * Coordinated the alignment of technology components and Deloitte's internal resources with overall project structures, ensuring effective integration and operation. * Oversaw the complete delivery of more than six proof-of-concept (POC) projects, implementing solutions on multiple platforms like Azure, Google Cloud Platform (GCP), and Amazon Web Services (AWS). * Created innovative operational solutions for LLMs using Azure PromptFlow, AWS Bedrock, and Google Vertex AI, which improved model management and deployment procedures. * Employed Langchain and llamaIndex frameworks to develop sophisticated language and diffusion model solutions, enhancing model performance and usability. * Performed thorough project estimations and risk evaluations, and developed strategies to mitigate risks to ensure seamless execution and delivery of technology solutions. * Conducted over 20 significant client meetings, effectively explaining complex AI solutions and architectures during client presentations and capability reviews. * Served as a vital link between the sales and delivery teams, maintaining a unified strategy and clear communication throughout all project stages. * Wrote impactful client-facing proposals, outlining the value, methodology, technology choices, and detailed cost-benefit analyses to support business development and client engagement efforts.

In the Claim Litigation project, we integrated Graph Databases and Natural Language Processing (NLP) tools via Google Cloud Platform (GCP) to manage and predict litigation claims. This solution entailed developing a sophisticated data pipeline to process claim data, utilizing machine learning for predictive analytics, and aiding effective decision-making. This implementation empowered the insurance firm to anticipate, manage, and reduce litigation costs. The Producer Segmentation project focused on categorizing insurance producers using Snowflake and AWS for data handling. We developed data pipelines to manage producer data efficiently and used machine learning for the segmentation process. This categorization, based on performance and other metrics, enabled the firm to tailor their support and marketing strategies effectively, enhancing overall productivity. The Premium Leakage project involved creating a comprehensive solution to detect and mitigate premium leakage. Leveraging the power of GCP and AWS for data engineering, we built data pipelines to handle and analyze large volumes of insurance data. Our use of advanced machine learning enabled the detection of patterns indicative of premium leakage. This approach helped the insurance firm address those issues, resulting in revenue enhancement.

Delivered structured training programs to 200+ students and professionals on core and advanced topics in Python, Data Analysis, and Generative AI. Curriculum covered: Python Programming: Data structures, OOP, file handling, web scraping, automation (with requests, BeautifulSoup, Selenium) Data Analysis: NumPy, Pandas, Matplotlib, Seaborn, data cleaning, EDA, feature engineering, and statistical analysis Machine Learning: Supervised/unsupervised models using Scikit-learn, XGBoost, and LightGBM Generative AI: Prompt Engineering: Zero-shot, few-shot techniques with OpenAI GPT-4 RAG Pipelines: LangChain, LlamaIndex, Pinecone, ChromaDB LLM Fine-Tuning: LoRA, PEFT using HuggingFace & transformers Model Evaluation: RAGAS, BLEU/ROUGE metrics, hallucination mitigation Deployment: Streamlit, Gradio, FastAPI + cloud (GCP/AWS/Azure) Project-based Learning: Led hands-on projects (LLM-powered chatbots, resume screeners, CV-based document classifiers, customer churn prediction) Career Outcomes: Enabled 50+ professionals to switch careers into AI roles and 100+ students to qualify for internships or full-time tech roles