Shibla TP

Developed and deployed AI-driven B2B lead enrichment pipelines using Python, GPT(LLM), and Apollo API to automate profile summarization and data storage in
Airtable/PostgreSQL.
Collaborating with cross-functional teams to integrate AI solutions into existing products and services
Developed OCR-based document extraction with Streamlit and Tesseract, integrated RAG AI chatbot using Pinecone for natural language queries.
Conducting research on the latest AI, machine learning, and data science technologies
and methodologies
Designed and implemented Flask-based APIs to power influencer scoring systems, including data extraction (Apify), engagement rate calculation, and GPT-based
influencer classification.
Built AI analyzers for social media bios, leveraging GPT to assess tone, audience alignment, and brand fit — supporting influencer marketing decisions.
Integrated cloud services (PythonAnywhere, PostgreSQL) with automation tools (Zapier, Make.com) to streamline CRM workflows across Calendly, Systeme.io, and
Airtable.
Created end-to-end marketing dashboards using PHP and Python, visualizing campaign performance data from LinkedIn and WhatsApp channels.
Applied GPT and LLMs for summarization, categorization, and decision-making in data workflows, improving marketing efficiency and targeting.
Collaborated across technical and marketing teams to deliver AI solutions that enhanced lead qualification, campaign targeting, and CRM intelligence.
Built Retrieval-Augmented Generation (RAG) pipelines to fetch contextual client insights, assess lead quality using custom scoring algorithms, and generate hyper-
personalized outreach messages aligned with each prospect’s business needs
Optimizing AI models for performance and scalability
Develop and implement statistical models, machine learning algorithms, and predictive models to solve business problems
Developed and deployed LangChain-powered automation workflows for marketing operations, combining chains, agents, tools, memory, and retrievers to enable intelligent data processing,

Designed and developed a web application to manage laundry services, including order tracking, customer management, pricing modules, and delivery scheduling.
Implemented user authentication, admin dashboards, and status updates for efficient order processing and service tracking.
Integrated relational databases (SQLite/PostgreSQL) to store customer and transaction data, ensuring scalability and data integrity.
Used Bootstrap for a responsive front-end interface, providing a clean user experience across devices.
Participate in the entire application lifecycle, focusing on coding and debugging
Advanced proficiency in Python and related technologies
Experience with web frameworks such as Flask or Django
Design, build and maintain back-end services and API

Designed and implemented a real-time object detection system using YOLOv8 and OpenCV to identify smartphones in video/image frames for use cases such
as exam monitoring, restricted zones, and vehicle safety compliance.
Optimized detection accuracy and speed through model tuning, data augmentation, and performance benchmarking under varied environmental conditions.
Built a news recommendation system using NLP techniques (TF-IDF, BERT) and libraries like Scikit-learn, TensorFlow, and Transformers to personalize article
suggestions based on user reading behavior and click history.
Applied both traditional ML and deep learning models for scalable deployment, ensuring adaptability to real-world dynamic content environments.
Developed and implemented deep learning models using python
Implemented image processing projects.
Developed web application using python frameworks.

Developed and implemented machine learning and deep learning models to automate the extraction of handwritten text from images.
Applied image preprocessing techniques (e.g., noise reduction, thresholding) to enhance text readability.
Evaluated model performance using accuracy, precision, and recall metrics to ensure high extraction quality.
Built a pipeline for image input, model inference, and text output to streamline real-time processing.