Email:sayiqajabeen14@gmail.

com
LinkedIn: linkedin.com/in/sayiqa-jabeen-machine-learning-engineer
Github: github.com/Sayiqajabeen
PHONE: +92 322 4478453
Location: Islamabad/Lahore

SAYIQA JABEEN

Artificial Intelligence/Machine & Deep Learning Engineer

SUMMARY
I have completed a specialization in Machine Learning Engineering, where I gained a solid foundation in
various machine learning algorithms and techniques. This education has equipped me with the necessary
skills to effectively analyze data and develop innovative solutions using machine learning models.
Furthermore, I have acquired hands-on experience in implementing models through my previous
experiences. My proficiency in Python, Pytorch, Tensorflow, and other relevant frameworks has enabled
me to deliver high-performance models that have significantly improved the accuracy and efficiency of
various projects.

EMPLOYMENT HISTORY

Artificizen (Present) AI/Machine Learning Engineer

Deep Learning Engineer

SmartHive Marketing (January-May 2024)
Machine Learning Engineer
NexusBerry (June 2023 - December 2024)

Napollo Software Design LLC (March-July 2023) Machine Learning Engineer

Internship at Prodigy InfoTech Machine Learning

Internship at CodeAlpha (July 2024) Machine Learning

TECHNICAL SKILLS
AI Tools : Development and Training, Data Processing and Manipulation, Visualization,
Processing, Model Deployment and Serving, Data evaluation, Django,
, Knowledge of cloud platforms AWS, Cloud, Azure.

Expertise : Problem Solving, Python, Machine Learning, NLP, LLM, Deep Learning,
AI skills: Machine Learning Frameworks, TensorFlow, Matplotlib, SpaCy, RAG,
Database, Neural Network(NN) , Data visualization, Scikit-learn, PyTorch,
Seaborn ,Dialogflow.

Programming Languages : Python, Javascript, HTML, CSS

 PROFESSIONAL EXPERIENCE
AI/ML Engineer Artificizen (Present)

 Simplified interface for building and training deep learning models.

 Supports Speech recognition, RETNET, RAG Design Chabot for IT service management.
 Seamless integration with other deep learning frameworks.
 Robust code editing with syntax, code completion, and error handling.
 Pre-training, fine-tuning, Alignment and post-training LLM.
 Working as AI expert in Interview Panel.
 Working as a Peer guide and Trainer for newly joiners.
Projects:
Medical Text Summarization using RETNET & Transformers
This project compares the summaries generated between the Transformer architecture and the Retentive Network
(RetNet) Model, which is a successor to Transformer for Large Language Modles. The RetNet model is trained on the
subset m2 from the dataset MSLR2022 by Allen AI Institute. The summaries are evaluated on the ROUGE and BLEU score
metrics.
Financial Data Processing and RAG System:
This project combines advanced financial data processing, analysis, and a secure Retrieval-Augmented Generation (RAG)
system to deliver a comprehensive platform. The system provides capabilities such as generating detailed graphs, performing
advanced calculations, and retrieving user-specific financial data securely. Designed with scalability and user-friendliness in
mind, the platform is tailored to meet modern analytical and operational requirements.
Speech App:
Speech to text and Image using transformers implementation through APIs/models & deploymet on huggingface gradio.
AI-Agent for Youtube Content Summarization & Exploration and Integration with Education Management tool:
Frontend with gradio for building an interactive user interface with components like textboxes, buttons, and data tables.
Integration with APIs youtube v3.

AI Expert at SmartHive Marketing (January –May 2024)

Projects:
Text Generation with a Pre-trained LLM GPT:
Using this model to generate different texts, analyze the generated text in terms of coherence,
creativity, and relevance to the prompt.

Face Recognition System:

Build a simple facial recognition system using OpenCV and a pre-trained FaceNet model. Use OpenCV to
capture images from a webcam. Implement facial detection and extract facial embeddings. Compare
embeddings to recognize known faces in real-time. Display the recognized names on the video stream.

Online kernal learning for active sensor network with 92% accuracy:
Integrate online learning with a simple active learning strategy. Basic sensor data stream and implement
a kernelized learning model using the scikit-learn library. For fitting and predication use SDG & RBF
kernel. Applying Simulating Sensor Data & Uncertainty Sampling 7 got 92% accuracy.

Bitcoin Price Prediction/ Trend Analysis (Currently working on this project):

Using historical data (2018-2024) applying LSTM model give the best prediction of future trend.
 Machine Learning Engineer at NexusBerry (June 2023 - December2023)
 Reviewing and understanding of Project requirement specification documents.
 Determine the objective of the model (e.g. classification, regression, clustering)
 Gather and Organize data from database, public dataset.
 Use visualization tools and statistical methods to gain insights.
 Perform data cleaning, feature engineering, normalization, splitting data.
 Designed a predictive model to estimation.
 Participating in ML meetings and discussion sessions.
 Able to work on different projects simultaneously. Can also work independently.
 Train, evaluate, testing, deploy, monitor and maintain model.
 Create visualizations (e.g., histograms, scatter plots, box plots) to explore relationships
between features.
 Keep up with the latest advancements in ML research and technologies.

Projects:
NLP (Spam email filtering):
 Naïve Bayes Classifier for Spam email filtering project through NLP.
Creating a spam email filtering system using the Naïve Bayes classifier and NLP
techniques.
 IPhone Reviews NLP using NLTK| Accuracy 85%.
LSTM for Robot Arm Control:
LSTM (Long Short-Term Memory) networks can be highly effective for controlling a robot arm,
particularly due to their capability of handling time series data and learning from sequential
dependencies. LSTM network Design, training network, testing, validation, deployment, feedback loop
continuously monitor the performance of the LSTM controller and update the model as necessary.

Amazon Product Review Sentiment Analysis using RNN (Using NLP):

Using the Product Reviews dataset which has around 25000 customer reviews.
The end goal is to give a rating according to the review given
Feedback collection Bot/Chatbot using Dialogflow Cloud:
A Chatbot designed to collect user feedback and surveys.
 Gathers user opinions and experiences efficiently.
 Analyzes feedback to identify trends and areas for improvement.
 Enhances user engagement by making feedback collection more interactive.
Uses sentiment analysis to classify feedback as positive, negative, or neutral and ML algorithms to
detect patterns and insights from the data.

Machine Learning at Napollo Software Design LLC (March-July 2023)

 Lead the design, development, and deployment of machine learning models to solve specific business
problems.
 Evaluate and choose appropriate algorithms and methodologies for different projects.
 Review the code developed by the ML team to ensure quality and adherence to best practices.
 Hire, train, and mentor ML engineers and data scientists, fostering a collaborative and innovative team
culture.
 Define project goals, timelines, and deliverables in collaboration with stakeholders.
Projects:
Forecasting currency fluctuations (Predictions):
 The Leveraged historical financial data and advanced predictive modeling
techniques to develop a robust model for forecasting currency fluctuations,
achieving an average margin of error of only 2%. The model incorporated various
economic indicators and market trends to enhance its predictive accuracy.
  Conducted extensive data preprocessing and feature engineering to ensure the
quality and relevance of the input data, significantly improving the model's
performance. Regularly validated the model using cross-validation and back testing
methods to ensure its reliability and effectiveness in real world scenarios.

ACADEMIC PROJECTS
 Residual Power Series Method by Obstacle Problem, Algorithms of Machine Learning.
(Final Year Project) at FAST-NUCES.
 Collaborated with cross-functional teams to integrate machine learning models into existing systems,
automating the process of matching predicted values with actual values.
 Reduced manual workload by 75% and increased prediction accuracy to 95% through effective
automation and integration.
 LSTM model for a time series using sine wave dataset:
Build and train an LSTM model for a time series prediction task using TensorFlow and
Keras with help of sine wave dataset.
 The Legend of Sleepy Hollow:
Cleaning Data for LLMs, from cleaning to vectorizing it applying NLTK and SpaCy.
 Classification of images through Convolution Neural Network model (CNN):
Using CNN CIFAR10 dataset for classing images across 10 categories along with Tensorflow.
 Flight fare prediction with 90% accuracy:
Draw some predictions about the price of the flight based on some features such as what type
of airline, arrival time departure time, duration of the flight, source, destination and more .
 Yearly time spent on a app with 90% accuracy
Applying linear regression and seaborn for visualization.

Education

MS Math with specialization in Machine Learning National University of Computer and Emerging sciences Lahore
(FAST-NUCES)
BS Math University of Education Lahore.

Awards:
Received Recognition winner at Napollo Software Design LLC
Received a Training Hero award at Nexusberry
Received appreciation award from SmartHive.

Certification & Skills

Interest:
 Certification from LinkedIn of GenerativeAI
Hiking
(https://lnkd.in/g3_7w2yY)
 Artificial Intelligence & Deep Learning certifications from Photography
Udemy (https://www.udemy.com/certificate/UC- Travelling
844c3af9- 673a-4ce0-937a-e4ac2cfc7062/)
 Data Visualisation: TATA Group
(MyXvBcppsW2FkNYCX_Tata
Group_J53reiCRjsaha6Sox_1723479763
300_completion_certificate.pdf (forage-
uploads-prod.s3.amazonaws.com))

 AWS Certified Machine Learning from

LinkedIn (https://lnkd.in/gqd6TpMe)