nagapavan525 Goto Github PK

To say about myself -- Have strong analytical(Stats/ML/DL/AIML), AIML products/solutions architect, project management, & problem-solving skills which aid in delivering AIML solutions that help my organization.

I am an AIML Software Product Manager or Delivery Lead at Accenture, for CPG.AI analytical workstream.

I help to build & deliver industrialized analytical solutions for CPG/FMCG clients using a combination of Statistical/ML/DL/AI modeling tools, Cloud Infra, Figma UI/UX designs, & UI/UX development which helps for designing the best business strategies in terms of marketing, product manufacturing, & other business functions' strategies. I work with multiple functional teams to deliver the end-end AIML solutions including Accenture's leadership, my AIML core team, Data provisioning teams, Data vendors/source teams, UI/UX design teams, UI development teams.

Under my leadership, we delivered multiple analytical solutions - Consumer infinity in March 2022 which made a big impact on Accenture Business & many clients and increased the revenue by 10% on average for clients after using it. The solution is basically built using Consumer reviews & Social media data which gives insights of Consumers (Consumer segments, consumer preferences & sentiments, what they are searching for) & Products (What features of the products are the most important, Brand analysis, Competitor analysis, & sentiments across) in 360 degrees which help CPG industries to make data-driven decisions accurately & improve their business strategies.

And delivered another solution called SocialAI which has major features of user journey & trendspotting.

And we are currently developing solutions for DTC (Direct to consumer) Clients.

In Accenture, I wear multiple hats of roles including Delivery Lead, AIMLArchitect, project manager, technical product manager, & responsible to deliver end-end AIML solutions for various business functions(Marketing, Commerce, ConsumerInsights) in CPG/FMCG companies

I was a ComputerVision Practitioner & DeepLearning engineer who worked on building various Deep Learning solutions in the image, text, & digital signal processing domains for IBM Watson. In the past, I worked as a Core Python backend developer for products IBM CloudPakSystem(PaaS), Seagate's ClusterStor HPC StorageSolution, Buffalo NAS product, EMC Vplex, Actifio CDS

But I'm also working personally on building AI Core computer vision based projects using pytorch framework.

I had developed quite a number AI projects including some major projects Chest x-ray report Generation, Phenomonia bounding box prediction, COVID-19 chest x-ray image classfication, Blood Pressure Prediction using PPG signals.

I can help collaborating with industries who need consulting on building Computer Vision based projects.

My DEEP LEARNING SKILLS:

1. Handle various types of data including tabular data, images, speech, text, digital signals using Deep Learning
2. DNN and CNN Concepts - Convolutions, Pooling Operations & Channels, Kernels, Activations, and Layers
3. A step-by-step way to build the best deep learning network architectures with one variable change approach in an efficient way.
4. Receptive Fields - The CORE fundamental concept behind computer vision to design an efficient network.
5. Batch Normalization, Kernels & Regularization and Mathematics behind them
6. Backpropagation and Advanced Convolutions - Depthwise, Pixel Shuffle, Dilated, Transpose convolution.
7. Advanced Image Augmentation Techniques - Albumentations, Richman's data augmentation, and benchmarks
8. DNN Interpretability - Class Activation Maps, the most powerful debugging tool at your disposal
9. SuperConvergence: Cyclic Learning Rates, One Cycle Policy
10. ResNets: Trained ResNet for TinyImageNet from scratch
11. YoloV2/YoloV3: Understanding YOLOV2/V3 Loss Function, Implementing Object Detection Training & Transfer Learning on Custom Objects from scratch (Pipeline - Data Collection, Annotating, Data Pre-processing, Image augmentation, Customizing the YOLO architecture as per the need, Model training, Model inference [Extract the frames of the video, Inference the objects in the frames, stitch the frames back to form video along with Audio])
12. Advanced Training Concepts - Optimizers, LR Schedules, & Loss Functions
13. The intuition behind state of art networks VGG16/VGG19, ResNets, Inception, RCNN family (RCNN, Fast-RCNN, FasterRCNN), YOLO v1/v2/v3, and Squeeze & Excitation network.
14. Sequence Models - RNNs, LSTMs/GRUs, and Bidirectional & Attention-based LSTMs/GRUs.
15. Image Captioning - Image Captioning - Integrating CNN + LSTMs
16. Monocular Depth Estimation and 3D plane surface prediction from 2D data which can be used in VR, AR, and Autonomous driving.

How to reach me or know about me:

• You can reach me at linkedin - https://www.linkedin.com/in/naga-pavan-kumar-kalepu-8b516b29/
• Email - [email protected]
• Phone/Whats app - +919948613631
• Resume - https://drive.google.com/file/d/1r6C0RH3QNwelVO9j6kC1Om3OiQ7xbsxj/view?usp=sharing
• Github Link - https://github.com/nagapavan525/
• AIML Project portfolio - https://eportfolio.greatlearning.in/naga-pavan-kumar-kalepu