Bhanu Venkat Srikakulapu

Data Scientist specializing in SQL and Python, with a focus on data visualization and strategic insights. Proven track record of enhancing operational efficiency through data-driven decision-making and collaboration. Skilled in developing GenAI-powered analytics pipelines that drive continuous improvement and support business growth.

Project I

AI Model Deployment on Renesas V2H Board :

• Deployed AI models from the NVIDIA NGC Catalog, including FasterViT, Segformer, DINO, Deformable DETR, and Mask2Former, by adapting them for the Renesas V2H board using the NVIDIA TAO Toolkit.
• Retrained pre-trained models (e.g., .pth files) with custom datasets and optimized them through pruning, quantization, and conversion to ONNX format for efficient edge deployment.
• Streamlined the deployment pipeline by translating models into V2H-compatible formats, enabling seamless testing and validation on the Renesas V2H hardware.
• Leveraged TensorRT and NVIDIA TAO Toolkit for model optimization, ensuring high-performance inference on edge devices

Project II

Multi-Format Document Processing and Retrieval System :

• Developed a semantic search and retrieval system using FAISS, Salesforce CodeT5, and MiniLM embeddings, storing them in ChromaDB, enabling efficient indexing and retrieval of multi-format documents
• Built a retrieval-augmented generation (RAG) pipeline with Ollama, Llama 3.1, and CrossEncoder, generating C code responses based on context retrieved from vector stores.
• Optimized document chunking, embedding generation, FAISS indexing, GPU-accelerated embeddings, and batch processing, improving scalability and inference speed.

Project III

Detect COVID-19 Using Check X-ray :

• Developed a deep learning-based web application using a CNN model to detect COVID-19 from chest X-ray images.
• Tackled challenges such as limited data, overfitting, and model sensitivity by applying data augmentation and regularization techniques.
• Integrated the trained model with a Flask backend to deliver real-time predictions via a user-friendly web interface.
• Utilized technologies like Python, Pandas, NumPy, Keras, Scikit-learn, Flask, HTML, CSS, and JavaScript

Project IV 

Customer Churn Prediction Using Machine Learning: 

• Built a predictive churn model using ML algorithms (Logistic Regression, Random Forest, SVM, KNN, XGBoost) in Python.
• Performed data preprocessing with Pandas, NumPy, StandardScaler, VIF; applied PCA and GridSearchCV for tuning.
• Visualized churn patterns using Matplotlib and Seaborn; evaluated models using accuracy, precision, recall, F1-score, and ROC-AUC.
• Implemented neural networks in TensorFlow/Keras with EarlyStopping and ModelCheckpoint to enhance performance and reduce overfitting

Project V

KPI Dashboards for Sales & Operational Teams Performance:  

• Designed and developed dynamic KPI dashboards in Power BI and Excel to track sales and operational team performance on weekly and monthly intervals.
• Implemented ELT data pipelines using Python, SQL, and Apache Airflow to extract, clean, and load performance data from multiple internal sources and spreadsheets.
• Automated data refresh and scheduling workflows with Airflow DAGs, ensuring up-to-date insights and high data reliability for stakeholders.
• Conducted Root Cause Analysis (RCA) and applied Lean Six Sigma techniques to identify underperformance trends and recommend process improvements.
• Utilized Power BI, Excel, Python (Pandas, NumPy, Matplotlib, Seaborn), and SQL for data analysis, visualization, and insight generation.

• Advanced Data Analysis with Generative AI - Coursera
• Certification in Machine Learning and Deep Learning - Udemy

Soft skills: problem solving, critical thinking, teamwork, time management, interpersonal communication