Home

munikoteswara - Data science

[email protected]

Location: Sacramento, California, USA

Relocation:

Visa:

MUNI KOTESWARA GENERATIVE AI - LARGE LANGUAGE MODELS - DATA SCIENCE Data scientist with a foundational analytical background, enriched by a recent focus on generative AI. Committed to leveraging data for insightful decisions and driving AI innovation. Ready to bring my blend of experience and passion for continuous learning to a team dedicated to advancing data-driven technologies. TECHNICAL SKILLS o Programming Language & Tools: Python, SAS E-Miner, NumPy, Pandas, Streamlit, Flask o Generative AI and Frameworks: Gen AI, Llama 2, OpenAI, Falcon, Mistral, Huggingface Transformers, LangChain, LlamaIndex o Vector Embedding Models: Sentence Transformers, text-embedding-ada, Huggingface o Databases: Vector Database, ChromaDB, Milvus, Pinecone, Weaviate o DevOps & Deployment: AWS, Azure, Docker, Kubernetes, SAS Visual Analytics, GCP including EC2, VMs, Sagemaker, Vertex AI, AML, ACR, ACI, Compute Engines etc o Visualization Tools: Power BI, Matplotlib, Seaborn, o Deep and Machine Learning Algorithms and Packages: Linear Regression, Logistic Regression, k-Nearest Neighbors (k-NN), k-Means, Decision Tree, TensorFlow, PyTorch, Ensemble Methods (Random Forest, Gradient Boosting, Bagging, AdaBoost), Dimensionality Reduction (PCA, Factor Analysis, Matrix Factorization), Recommendation Systems, Support Vector Machines (SVM), , SciPy, OpenCV, Theano, Keras, Spacy EDUCATION Master s Degree (Master of Computer Applications) Jawaharlal Nehru Technological University, India Bachelor s Degree (Statistics & Mathematics) Andhra University, India PROFILE SUMMARY o Generative AI Expertise: Proven track record in developing and scaling AI solutions using advanced frameworks like LangChain and LlamaIndex, driving innovation in generative technologies. o Database Optimization: In-depth experience in managing and optimizing vector databases such as ChromaDB and Pinecone, significantly enhancing data retrieval processes for AI applications. o LLM Customization: Competent in implementing and fine-tuning a variety of LLM models (open-source and proprietary), tailoring generative solutions to align with specific project needs and performance benchmarks. o Cloud Deployment: Demonstrated proficiency in utilizing AWS & Azure services along with on-prem infra for the deployment of AI models, ensuring scalability and reliability in cloud services for complex AI operations. o Project Leadership: Strong leadership in cross-functional teams, delivering data-driven solutions and insights that have consistently surpassed strategic goals in fast-paced environments. EXPERIENCE Data Scientist, Gen AI Centene Corporation, USA Oct 2023 Till Date o Collaborated with diverse business units and stakeholders to elicit requirements, develop initial LLM model blueprints, and refine the proposed design in collaboration with architects. o Engineered and implemented an advanced customer service chatbot utilizing Large Language Models (LLMs) to enhance user engagement and streamline support functions. o Optimized predictive model accuracy by meticulously adjusting prompt engineering, chunking size, experimenting different Embedding techniques o Spearheaded the enhancement of a Retrieval-Augmented Generation RAG model by implementing advanced re-ranking algorithms, significantly improving response accuracy and relevance in complex query scenarios o Analyzed the business, its products, customer base, and market landscape to plan a cost-effective and secure LLM model that facilitates user interaction o Develop the model in Pycharm & test prior to creating docker images. o Create docker images & push to respective cloud platform like Az Container Registry or Google Artifact Registry o Tag the images to AZ App services, or VMs for compute in Azure or App Engines in google and AWS ELastic Beanstalk. o Automated ML lifecycle with integrated models in Azure DevOps for version control, CI/CD, and deployment. o o Closely work with Annotators to produce the Annotations on the selected data which is crucial to evaluate the output accuracy during the model build & testing. o Work with DevOps teams on the model deployment, monitor the deployment & performance. o Regular interactions with the business teams on the impact of newly deployed products to monitor the model efficiency. Data Scientist Digi Telecommunications Sdn Bhd, HCL, India Mar 2021 - Sep 2023 o Optimized predictive model accuracy by meticulously adjusting hyperparameters, resulting in a significant improvement in the F1 score. o Addressed the challenged outputs with various experiments using different text splitters, chunking techniques, RAG methodologies & prompt templates etc o Fine tune the RAG outputs with the help of RAGAS, Reranking, Recall techniques. o Experimented additional Embedding models for vector embedding including Cohere, OpenAI ada, HuggingFace as part of streamlining the overall pipeline. o Based on input data, trial, test & evaluate which chunking strategies works well, which embedded model could give better semantic indexing & how to manipulate the output using different ways of prompt templates, RAG architecture. o Research on what other models are available in the market with respect to security, pricing, accuracy & efficiency etc & provide the analysis to business units based on their requirements & budgets. o Coordinate with multiple departments to collect relevant historical data, work with Data engineers for data processing & sit with stakeholders to finalize the business requirements. o Engineering and deploying a potent Deep Neural Network using Python to forecast customer churn, fortifying customer retention efforts. This would involve using Python's rich ecosystem of machine learning libraries like Keras or Scikit-Learn. Maintained a centralized model registry for versioning and sharing machine learning models. o Built data-driven solutions leveraging ML, Generative AI (Llama 2), and Deep Learning to tackle business challenges. o Streamlined large-scale data analysis via integrated ML models in Azure Databricks o Optimized NLP workflows by integrating re-ranking mechanisms into RAG models, achieving a marked increase in precision for dynamically generated content across diverse datasets. o Spearheaded the enhancement of a Retrieval-Augmented Generation model by implementing advanced re-ranking algorithms, significantly improving response accuracy and relevance in complex query scenarios o Automated ML lifecycle with integrated models in Azure DevOps for version control, CI, and deployment. o Evaluate the RAG performance using metrics including Accuracy, Faithfulness, Answer Relevancy etc for generation quality. o Utilized AWS sagemaker and titian foundation models to build, train, and deploy machine learning and LLM models at scale. o Closely work with Annotators to produce the Annotations on the selected data which is crucial to evaluate the output accuracy during the model build. o Work with DevOps teams on the model deployment, monitor the deployment & performance. o Regular interactions with the business team on the impact of newly deployed products to monitor the model efficiency. o Identified and rectified deficiencies in existing business protocols by applying predictive analytics to strengthen churn detection capabilities. o Analyzed claims data to uncover diverse churn scenarios, including geo-location patterns, location-specific frauds, duplicate billing, upcoding, and excessive modifier usage. o used clustering techniques to classify based on their usage, location, subscription packages o Optimized the output by working on weights & Biases, back propagation & epochs. Data Science, Associate Consultant Manulife Insurance, DXC Technology, Malaysia Dec 2018 Feb 2021 o Work with Data engineers to collect and organize datasets, including images, and metadata. o Coordinate with engineers for data clean and preprocess data using Python libraries like Pandas and NumPy for normalization, handling missing data o Implemented an automated extraction and analysis of key details from accident reports, streamlining the claims verification process and improving the accuracy o work with the Architect on the model design & associate with the team on the model development with respective algorithms like CNN & YOLO. o Deploy the model in cloud based service with DevOps team on the services including AWS S3, EC2, Lambda, IAM, CloudWatch o Maintain comprehensive documentation of the development process & manage code versions and collaborate using GitHub. o Monitor model performance using Amazon CloudWatch, adjust models and algorithms based on performance data & retraining the procedures. o Reduced fraudulent claims by 20% through automated image analysis of accident scenes, leading to significant cost savings for the company. o Enhanced processing speed by 30%, improving claim resolution times and customer satisfaction. Senior System Specialist Dnata, DXC Technology, Malaysia Aug 2018 Dec 2018 o CAT1 tickets to work on Wintel related tasks o Plan cluster VM migrations to avoid concurrent moves of active & passive nodes. o Performed ESXI host patching from VMware update manager. o Deployed VMware Horizon VDI Instances and creating over 10,000+ desktops by building an entirely new infrastructure of Horizon View 7.x also integrated with existing Infrastructure o Installed and configured VMware Horizon VDI View Connection Servers o Image life cycle management for VDI deployment in the infrastructure with respect to project & departments o Supported VMware farm hosted on Nutanix hyperconverged platform thru Prism Element/Central. o Troubleshoot issues over ESXi hosts, VCSA thru PuTTy CLi and also HW console for HP & Dell etc o vSwitch, port groups, Storage, cluster HA & DRS configurations Infra Project Engineer Visa Worldwide, Geco, Singapore Jun 2017 Mar 2018 o Implemented Zerto tool for on-prem to cloud migration project & completed the project earlier than project timeline. o Weekly decks presentation to customer on capacity for existing infrastructure including Compute & storage for both cloud & non-cloud platforms. o Quarterly security pathing for hosts using LCM & addressing the vulnerabilities identified by periodic scanning of servers. o Configuring and managing ESXi Clusters with HA Settings and Customizing HA for Virtual Machines. o Work with datacenter field engineers for new server setup & installations & coordinate for Hardware related issues along with Vendor support. o Upgrade HP iLO, Dell iDRAC FW & HW components including array, NW, HBA card etc from respective vendor bundles followed by Server OS version & Firmware. o Based on business requests, convert physical servers to virtual (P2V) & V2V for case to case requirements. o Perform DR/BCP along with SAN replication, Network, applications, DB, App servers failover etc. o Support GPO Group Policies, AD schema, Organization Units (OU) etc Project Engineer Schroders Investment Management, Robert Walters, Singapore Jan 2015 Jun 2017 o Plan, design new architecture where necessary to enhance existing vSphere VC infrastructure. o Overall responsible for end-to-end project related tasks mainly in technical areas where to find out legacy systems, analyze the criticality, dependencies, negotiate with stakeholders, application & server owners on the design, schedule, capacity planning of new HW & in Virtual platform. o Setup new HW in datacenter, configure RAID, Install ESXi & vCenter servers, configure cluster HA, DRS & deploy VMs with respective guest OS. o Managing multiple vCenters with over 2000 esxi hosts in various regions hosting thousands of virtual machines including critical applications, databases, web servers etc. o Installing, configuring & managing Virtual Center server. Experience on configuring data stores to ESX server using SCSI, iSCSI or LUN. o Expertise in installation, configuration, and administration of Windows and VMWare servers o Cluster configurations & maintenance for Windows, ESXi, SQL, File servers etc. o Orchestrate ESXi host patch updates via Lifecycle manager o Monitor vSphere farm using vROPs alerts & highlight issues identified to management for pro-active measurements. o vCenter cluster configuration for HA, DRS, vMotion, SvMotion, affinity rules to keep mission critical VMs availability. o Extract the WWPN & WWN from new hosts & existing hosts for additional/external storage provisioning. o Work with SAN team for Zoning w.r.t ESXi hosts & storage device controllers o Host grouping & storage LUNs provisioning based on cluster or standalone hosts requirements. o Configure datastores in ESXi hosts & clusters & plan for datastore migrations from EMC to Pure storage. o Troubleshoot vSphere related issues including vCenter, ESXi hosts & VMs. Collect detailed issue info from vROPs & fix the issues or raise a VMware support case accordingly. o Propose latest technology & automate tasks to reduce infra resources by saving cost, resources, manpower. o Analysis and negotiation on user requirements on the server resources for better capacity planning o Work with the PMO team on new hardware procurements. Involve in delivery, set up of HW in DC & support discussions with vendors System Engineer Hewlett Packard, MindTeck, Singapore Jul 2014 Dec 2014 Project Engineer CIMB Bank Berhad, Hanodale, Malaysia Jan 2012 Feb 2014 CERTIFICATIONS: Amazon Web Services Cloud Architect certified Microsoft Azure certified VMware vSphere VCP certified Keywords: continuous integration continuous deployment artificial intelligence machine learning business intelligence sthree database active directory rlang hewlett packard Arizona