Sk Shahid
Development
Odisha, India
Skills
Data Science
About
Sk Md Shahid's skills align with IT R&D Professionals (Information and Communication Technology). Sk also has skills associated with Programmers (Information and Communication Technology). Sk Md Shahid appears to be a low-to-mid level candidate, with 3 years of experience.
View more
Work Experience
Freelance Data
December 2023 - Present
- NLP Project - Clustering:Implemented Word2Vec and NLTK for text processing, transforming menu descriptions into numerical vectors, and used K-means clustering to categorize restaurants by menu profiles and pricing, offering valuable insights for owners to refine their offerings and pricing strategies. • Time Series Project:Executed a time series analysis by fitting ARIMA and SARIMA models, comparing their performances to derive optimal forecasting solutions.
Subject Matter Expert
PHY & MAT
September 2020 - November 2023
- • Led the team by offering online tutoring, and helping with assignments and Projects. • increased the team size from 2 to 6 SMEs by providing 90-95% average Grade scores. Notable TOPIC MODELLING W Projects Objective: Developed a hybrid model combining unsupervised clustering and semi-supervised methods to enhance tweet categorization using social media hashtags. Tools used: Python, Pandas, ML Algorithms, Word2Vec, PyLDAvis Roles and Responsibilities: • Implemented a semi-supervised learning framework to cluster short texts, employing Word2Vec for semantic analysis. • Optimized clustering accuracy using dimensionality reduction methods (t-SNE, SVD, PCA), ensuring pattern preservation and noise reduction. • Conducted comparative analyses of multiple clustering algorithms, including K-means, GSDMM, and fuzzy clustering, to determine optimal performance. AI TUTOR W Objective: Developed "AI Tutor - Acharya, " an innovative web application using cutting-edge LLM & RAG architecture to provide personalized tutoring in Hindi. Tools used: RAG, FastAPI, Langchain, Qdrant DB, Docker Roles and Responsibilities: • Compiled and processed a diverse range of educational content, including textbooks and online resources, to create a comprehensive database for model querying. • Built the backend of the system using RAG pipeline and FastAPI. • Deployed and continuously monitored the system's performance, focusing on adaptive learning experiences tailored to individual student needs. PHYSIGEN W Objective: Led a project to develop an advanced AI-based tool using Large Language Models (LLMs) for generating a dynamic question bank, addressing the scarcity of diverse, high-quality JEE Mains physics study materials and students' tailored learning. Tools used: Prompt Engineering, Web scrapping, OCR, Fine-Tuning (LoRA), Inference optimization, Gradio. Roles and Responsibilities: • Dataset Curation and Management: integrating advanced data gathering and processing techniques, including OCR, equation parsing, web scraping, and GPT-3.5-driven transformation. • Model Fine-Tuning: Utilized the Low-Rank Adaptation (LoRA) technique and optimization tools like WandB and Axolotl, effectively customizing and refining the AI model to accurately emulate the style and complexity of JEE Mains physics questions within resource constraints. • Model Inference and Application: Managed the model's inference process with precise prompts and integrated platforms like Bactrian X and Gradio for practical deployment, while overseeing initial evaluations to confirm its effectiveness and adaptability. Activities & Workshops and Conference Presentations: Acheievements • Participated in WAT 2023 workshop, Kyoto University and NICT, hosted by MT Summit 2023, Macau SAR China, September, 2023. • Presented at MT Summit 2023, Macau SAR China. • Presented OdiagenAI's work at Odia's in AI ML Global Conference, October, 2023. • Conducted Generative AI and LLM Workshop by OdiagenAI, November, 2023.
Education
Silicon Institute of Technology
Master of Science
January 2022 - Present
