🔭 I possess a strong enthusiasm for Artificial Intelligence, Deep Learning, Machine Learning, and Natural Language Processing. My professional background spans E-commerce, Finance, Supply Chain, and Product domains. My focus revolves around crafting ML/AI solutions from inception to deployment, with a keen interest in end-to-end implementation.
💭 My objective is to actively engage in mentoring startups, guiding them through the intricacies of AI implementation for the public good. I aspire to leverage cutting-edge AI technologies to develop impactful solutions that address complex societal issues and streamline everyday experiences for individuals globally. By harnessing the power of artificial intelligence, I seek to catalyze positive change and contribute to the advancement of society in meaningful and tangible ways.
👯 With over 100+ mentoring sessions on Data Science and AI, I'm also deeply engaged in Gen AI and Deep Learning research. Let's collaborate! Feel free to message me at email for research opportunities.
- 📝 I regularly write articles on https://medium.com/@shivamnegi92
- 📫 How to reach me negi.sh@northeastern.edu
- 👨🎓 Book a mentoring call https://topmate.io/shivam_negi_data
- Substack AI Core Loop
🚀 Currently, I serve as a Principal Data Scientist at Walmart Claim Services, leading the charge in integrating GenAI and data science methodologies to elevate our organizational landscape. My role involves orchestrating the multi-cloud deployment, auto-monitoring, and optimization of advanced AI/ML models, including applications of Large Language Models (LLMs)- (GPT4,LLama, FlanT5 etc) in projects spanning topic modeling, Retrieval Augmented Generation (RAG), summarization, risk detection, named entity recognition, prompt engineering. 🤖
⌛️ During my tenure at Amazon, I played a pivotal role in constructing end-to-end machine learning models tailored for supply chain optimization. Additionally, I led the development of analytical infrastructure for key Amazon ventures like Amazon Prime and Amazon Fresh, leveraging AWS cloud services 📊. My focus extended to meticulously tracking over 100 KPIs, operating within a startup-like environment that demanded agile thinking 🚀 and rapid adaptation to evolving challenges!
🛠️ Moreover, I spearheaded the implementation of multiple event-driven pipelines leveraging AWS Lambda, SageMaker, S3, and other key services. This approach enabled seamless data processing and model deployment, driving efficiency and innovation within our analytics ecosystem.
🎓 With a Masters in Information Systems from Northeastern University, Boston, and a Bachelors from VIT University, Vellore, I engaged in impactful United Nations Research Projects at Northeastern. Employing advanced data science and statistical techniques, I analyzed migration patterns in Eastern Europe, using geolocation and macroeconomic data to inform public policy decisions.
Feel welcome to connect with me on LinkedIn for the latest updates on my AI learning materials!
| Repo | Description |
|---|---|
| voice-agent-eval-corpus | Companion repo to Evaluating Real-Time Voice Agents (arXiv:2609.30798): a verified 41-source corpus and TRG, a minimum reporting standard for voice agent evaluation, with a live demo |
| tinymorph | Self-improving training data for small language models via temporal weak-strong curation - your model at epoch 1 becomes the weak oracle, no external APIs needed |
| Cancer-Image-Classification-ViT | Vision Transformer applied to cancer image classification |
| Repo | Description |
|---|---|
| awesome-entity-extraction | Curated, CI-verified list of LLM-era entity, relation, and structured-data extraction tools |
| awesome-multimodal-extraction | Curated, CI-verified list of multimodal (image + text) extraction tools: VLMs, document AI, OCR |
| practical-ai-systems | 30 offline-first, runnable AI engineering systems with tests, evaluations, and browser demos |
| yesman | Stops AI coding agents from empty praise and hype; works in Claude Code, Cursor, Codex, Copilot, Windsurf, Cline, and more |
| Repo | Description |
|---|---|
| AICave | Deep learning from scratch: Transformer, Llama, ViT/JEPA implementations plus PyTorch fundamentals notebooks |
| GenAI-Playlist | Fundamental GenAI projects built from scratch |
| RAG_END_TO_END | End-to-end RAG variations: chatbots, agentic RAG, Graph RAG |
| FastAI_Pytorch | fastai and PyTorch learning notebooks |
| DeepLearningPlaylist | A knowledge database covering the major types of deep learning models |
| Datawarehouse-and-Tableau-PowerBI | Business intelligence reports and dashboards on Tableau and Power BI |
| Big-Data-and-Intelligence-Analytics | IBM HR Analytics employee attrition analysis with Decision Trees and Keras |
| LSTM-RNN-over-Bitcoin-Historical-Data | LSTM-based deep net predicting 30-day Bitcoin value trends |
| LSTM-BLSTM-based-Image-Caption-Generator-leveraging-Inception-V3 | Image caption generation using LSTM, BLSTM, CNN, and InceptionV3 |
| SentimentPredictionML | Sentiment prediction with classical ML models |
| UC-Berkeley-CS188-Intro-to-AI--Project-1-Search-in-Pacman | Berkeley CS188 Intro to AI: search algorithms implemented in Pacman |
| Deep-learning-Keras-Model-over-Semeion-Image-Data | Keras deep net trained on handwritten digit image data |
| Prudential-Life-Insurance-Assessment--Machine-Learning | Predictive risk-classification model, classical ML approach |
| Prudential-Life-Insurance-Assessment--Neural-Net | Same risk-classification problem, neural net approach |
| World-Development-Indicator-SouthAsianCountries | Exploratory data analysis on development indicators across South Asian countries |
Full activity: github.com/shivamnegi92?tab=repositories
