Shaheen Nabi

Data Science Intern – iNeuron.ai

January 2025 - March 2025 | Bengaluru, Karnataka, India

• Developed a high-precision object detection model using YOLOv5 to identify and classify crops and plants for agricultural applications.
• Annotated and open-sourced a 25,000-image dataset on Hugging Face, advancing global agritech research.
• Designed and deployed an automated AI agents pipeline using AWS ECR, EC2, and Jenkins for real-time inference and insights.
• Trained models on NVIDIA A100 GPUs, optimizing for high performance and real-world deployment.
• Conducted rigorous testing across diverse agricultural environments to ensure model robustness.
• Integrated SMTP for automated delivery of personalized research reports via a user-friendly UI.
• Collaborated with teams to ensure seamless integration and scalability of the solution.

Founder – Lasso Pacific Pvt Ltd

January 2022 - December 2022 | Anantnag, Jammu & Kashmir, India

• Founded an AI-driven edtech platform to democratize AI and computer vision education, targeting rural students globally.
• Developed hands-on courses in AI, computer vision, and autonomous systems, making complex topics accessible.
• Attracted 2M+ annual visitors organically through high-value content and career roadmaps.
• Managed curriculum development, partnerships, and global outreach to maximize educational impact.
• Acquired an initial user base, validating demand for accessible AI education in underserved areas.
• Navigated funding, scaling, and growth challenges, gaining expertise in entrepreneurship and product development.
• Closed the venture but empowered thousands of learners with AI knowledge and career guidance.

🌾 LeafLogic Multi-AI Agents Project

Jan 2025 - Mar 2025

• Developed a production-ready, fully automated AI system capable of identifying 100+ plant and crop species with a one-click solution, ensuring fast and efficient detection.
• Designed and implemented a high-performance object detection pipeline leveraging YOLO, great for real-world agricultural applications.
• Built a Multi-AI Agents framework that autonomously retrieves enriched plant data from the web, extracting insights on species characteristics, optimal growth conditions, price details and commercial applications.
• Engineered a scalable and cloud-optimized infrastructure, integrating AWS ECR, EC2, enabling real-time inference and seamless deployment.
• Automated the complete AI pipeline, covering training, and post-detection research, significantly reducing manual intervention.
• Utilized OpenAI's Chat Completion API (GPT-3.5 Turbo) alongside Serper API, Exa API, and other external services, leveraging the TaskFlowAI open-source framework to build a robust Multi AI-Agents pipeline.
• Conducted extensive testing and validation in diverse agricultural environments, ensuring the model’s adaptability and high performance in real-world conditions.
• Ensured modularity and scalability, allowing easy expansion to support additional crop species and agricultural use cases in the future.
• Enabled customers to receive their summarized reports directly via email by entering their email address on the UI, ensuring a seamless and convenient user experience.

Tech Stack: Python, PyTorch, Jenkins, AWS, TaskFlowAI, OpenAI, SerperAPI, A100 GPUs, Hugging Face

🌍 TripPlanner Multi-AI Agents Project

Dec 2024 - Jan 2025

• Developed an intelligent travel assistant integrating three AI-driven agents to automate and optimize the travel planning experience.
• Web Research Agent: Gathers comprehensive destination insights, including local attractions, events, weather updates, and essential travel information.
• Travel Planner Agent: Creates personalized itineraries, suggests appropriate clothing based on real-time weather forecasts, and recommends optimized travel routes.
• Report Generator Agent: Utilizes GPT-3.5 Turbo to compile a concise and actionable travel report, consolidating all essential details for seamless trip planning.
• Designed for efficiency and personalization, ensuring travelers receive relevant, well-structured information tailored to their preferences and schedules.
• Built with modularity and scalability, allowing easy integration with additional travel services, AI enhancements, and real-time data sources.
• Optimized for travel enthusiasts, business professionals, and frequent flyers, providing an AI-powered solution for hassle-free and well-organized trips.

Tech Stack: Python, TaskFlowAI, OpenAI, GitHub Actions, ECR, EC2, Streamlit, SerperAPI

🧠 Instruction Fine-Tuning of Meta LLaMA 3.2 3B

Nov 2024 - Dec 2024

• Fine-tuned the Meta LLaMA 3.2-3B Instruct model using LoRA (Low-Rank Adaptation) and 4-bit quantization, enhancing its ability to generate context-aware responses in Kannada.
• Focused on improving the model’s proficiency in understanding and generating highly relevant, instruction-following responses in conversational inputs.
• Utilized the Hugging Face dataset charanhu/kannada-instruct-dataset-390k to refine contextual accuracy and linguistic fluency.
• Pushed the fine-tuned model to the Hugging Face Hub, where it is actively used by hundreds of developers monthly for Kannada NLP applications.
• Developed using the Unsloth framework and trained on Google Colab Tesla T4 GPU, ensuring an efficient fine-tuning process.
• Merged the base model and fine-tuned model into 16-bit floating point (FP16) for optimized deployment and inference.

Tech Stack: Python, Unsloth, QLoRA, Instruction Fine-Tuning, SFT, Hugging Face, Transformers, TRL, Quantization, AWS

Cyber Fraud Prevention Volunteer – Cybersecurity Outreach, Rural India (Self-Organized)

January 2020 - December 2023 | Social Services

• Initiated a grassroots campaign to educate rural Indian communities on cybersecurity threats and scam prevention.
• Conducted one-on-one sessions and workshops, teaching digital safety practices like phishing detection and two-factor authentication.
• Prevented numerous scams, saving victims tens of lakhs of rupees through education and intervention.
• Empowered individuals with practical digital literacy, fostering confidence and sparking interest in digital safety.

Senior Writer – Shexion

January 2020 - October 2020 | Education

• Volunteered as a cybersecurity content writer for Shexion.com, creating articles and roadmaps for aspiring cybersecurity professionals.
• Authored guides on networking, Linux, ethical hacking, and tools, plus analyses of cyber incidents like WannaCry and Pegasus.
• Reached tens of thousands of readers, providing resources that supported career starters in cybersecurity.
• Received feedback crediting content with clarity and career guidance, reinforcing my passion for tech education.

Writer – Lasso Pacific Pvt Ltd

March 2022 - October 2022 | Science and Technology

• Created educational content on Python, machine learning, computer vision, and autonomous systems for beginners.
• Authored career roadmaps to make AI accessible, especially for rural students with limited resources.
• Provided personalized career advice, encouraging young learners to pursue AI and tech careers.
• Contributed to Lasso Pacific’s 2M+ annual visitors with beginner-friendly AI and data science content.