Hello, I am Manoj Patil M a seasoned data scientist and AI enthusiast with over 1 year of extensive experience in the field of artificial intelligence, machine learning, and natural language processing. Throughout my professional journey, I have had the privilege of contributing to a diverse range of projects, each pushing the boundaries of technology and innovation. My expertise spans various domains, including generative AI, language modeling, fine - tuning LLM and deep learning models for specific use case, prompt engineering, chatbot development, classification tasks, supervised and unsupervised techniques, and projects. advanced analytics, and algorithmic solutions.
1)I spearheaded the development of a conversational Agri Bot for the Outgrow Farm Support App, equipping it with support for five Indian languages and the capability to accept both text and speech inputs. In this endeavor, I conducted prompt engineering, harnessing cutting-edge advancements in generative AI to enhance the chatbot's performance. Additionally, I enriched the bot's knowledge base with company-specific information and integrated an agent to retrieve answers from the internet. The deployment was seamlessly executed in an Azure Kubernetes environment using Docker. Through extensive fine-tuning efforts, I significantly improved the bot's response accuracy, achieving an impressive 80 percent accuracy rate.
2) Devloped leaf pest and disease classification model by implementing transfer learning techniques. Handling a substantial dataset of 600 gigabytes for model building and data preprocessing, I successfully enhanced the model's recall and precision, achieving an accuracy rate of 85 percent. Furthermore, I deployed the model using a CI/CD pipeline via Docker in an Azure Kubernetes environment, where I actively monitored its performance. As a result of these improvements, we observed a significant 15% increase in daily requests and received numerous positive feedback comments from satisfied customers.
3) I developed an analytics dashboard for our organization, utilizing a BI tool to facilitate data visualization and monitoring. This enabled us to provide valuable insights and reports to both clients and customers, tailored to their specific requirements. My collaboration with product managers and clients allowed us to extract meaningful insights from IoT data. Additionally, the implementation of this analytics solution resulted in a significant time savings of 15 hours per week.
4) Through the development and implementation of multiple algorithms for irrigation recommendation and pest and disease prediction, iot data monitor and implemented oops, DSA concept optimzed the algorithms by using memory mengement, threding and genertors. I significantly improved the accuracy of our solutions when tested with real-world data. As a result, we witnessed a remarkable increase of 35% in farmer subscriptions, demonstrating their heightened satisfaction with the results.
5) I have executed a variety of automation tasks, including web scraping from multiple websites, performing ETL operations, data preprocessing, Api developemnt,integrating with Jira, triggering emails, and generating and scheduling reports. resulted in significant time savings of 8 hours per week.
I possess a strong and diverse skill set in the field of data science, encompassing MLOps, large language models (LLM), prompt engineering, deep learning, classification, forecasting, and natural language processing (NLP). In addition to these core competencies, I also excel in data analytics, feature engineering, data visualization, statistical modeling time series analysis and model deployment. My comprehensive skill set empowers me to tackle a wide range of data-driven challenges and deliver impactful solutions in the ever-evolving field of data science
Gained an in-depth understanding of technical skills required for data science. Worked on real-world ML projects. Improved my story-telling skills by delivering presentations.
Creating 3d models and validating design, testing of CAD software's in eyeshot software. Developing nesting software in eyeshot, manual testing softwares,Creating 3d models and validating design, testing of CAD software's in eyeshot software. Designing products and considering project requirements (performances, manufacturability, serviceability, testability, etc.).
Secured 9.19 cgpa in my mtech
secured Fcd with 7.77 cgpa
Built conversational chatbot for farmers with multilanguage support with open source llm, have added internet search agent option to fetch answer from the internet if llm don't know answer or question related to the current trend. it will support input either text or speech.
Demo videothis project will support different types of input files all of which will be stored in an open source vector database. with open source llm or chatgpt we can multiple action with data, q&a, summarization
Demo videoBuilt a classification model to detect pests and diseases in crop leaf images. deployed it in Azure Kubernetes with docker.
Demo videoThis project will support csv,excell,pdf,docs,txt,ppt any type of file. it will extract data from the file based on data it will do summarize the data and generate mind map for the data
Demo video Github linkCrated research agent with open source llm, this agent can learn automatically form the older responses. i have provided internetscrapping support for get information from internet. based on any user query and selected agent type this will do researchon by using its core knowledge and internet .final ouput will be in structure format with intorduction, body and conclusion.
Demo videoThe main objective is to create a book recommendation system for users. based on Model Based Collaborative Filtering , Memory Based Collaborative Filtering , content based filtering, recommendation for new user with popularity
View Project View deployed ProjectGoal of project is Predicting energy consumption of home appliances . Did eda, feature selection data cleaning, feature engineering, implemented multiple regression algorithms did hyper parameter tuning and cross validation and model explainability.
View ProjectGoal of the project is predict weather customers is interested in vehicle insurance or not. Did eda, feature selection, feature engineering, data cleaning, imbalance techniques,implemented multiple classification models did hyper parameter tuning and cross validation and suggestion for the business improvement.
View Projectdeploy machine learning model using docker and flask. with flask created web api and using docker created image and container for model
View ProjectComplete eda on Telecom churn data set and analysis what are factors affecting customer churn and come up with business strategy
View Project