Asghar Mustafa

Asghar Mustafa

About

Detail

AI Strategist | AI Systems Architect | Business-aware Data Scientist | From ideas to impact: Helping SMBs save 10-20 hours weekly through Agentic AI & ML
Lahore, Punjab, Pakistan
Contact Asghar regarding: 
work
Full-time jobs
Starting at USD40/hour
Flexible work
Starting at USD45/hour

Timeline

work
Job
school
Education
folder
Project

Résumé

Jobs verified_user 0% verified
  • Lahore University of Management Sciences
    Adjunct Faculty - AI & LLMs
    Lahore University of Management Sciences
    Jan 2025 - Current (1 year 6 months)
    - Design and deliver course content on Large Languages Models (LLM) systems, AI Lifecycle and it's practical application in business. - Guide students in applying AI research and engineering practices to real-world challenges. - Mentor students on AI career pathways, industry expectations, and technical problem-solving. - Introduce best practices in scalability, optimisation, and responsible AI deployment.
  • freelance
    Senior Consultant (Solutions Architect) public Remote experience
    freelance
    Jul 2024 - Current (2 years)
    • Developed and deployed a fine tuned LLM using PyTorch to Azure, automating sentiment analysis for the Marketing Agency. • Collaborate with medical domain experts to improve medical processes using NLP & AI, for community treating cancer patients. • Worked on adapting ML pipeline to meet performance management of runtime and model optimization. • Deploy the running ML pipeline to Azure Container registry and Azure ML with software backend running through Azure Batch Processor.
  • C
    Senior Consultant (Assoc. Architect) – Data Science
    Confiz Limited
    Feb 2021 - Jun 2024 (3 years 5 months)
    • Lead the team to develop AI Solutions for Recommender Systems, Generative AI RAG and Agentic RAG use-cases using frameworks such as OpenAI, Huggingface, LlamaIndex & AutoGen with Azure cloud to materialize two new sales-leads for the business. • Conduct and document time series forecasting and anomaly detection using probabilistic, statistical and machine learning methods to provide intelligent forecasts on store retail and supply chain initiatives to meet the staffing requirements and streamline finance operations for the customer. • Delivered ML System Solution Design using GCP services including BigQuery, Google Cloud Storage & DataProc. • Collaborate and lead a team of DSs and DEs to successfully orchestrate PySpark ML pipeline to
  • M
    Data Science Tutor
    Monash Business School
    Jul 2019 - Feb 2021 (1 year 8 months)
    • Guided students on core concepts from statistical thinking and wild data analysis using excel and R. • Assisted candidates to develop statistical algorithms in machine learning to solve business problems. • Worked with data structures in Python & R to assist candidates to solve problems.
  • A
    Data Analyst – Technical Team Lead
    Afiniti Inc.
    Jul 2016 - Jan 2018 (1 year 7 months)
    • Automate service data logs analysis using Python regex to for high and critical incidents, resulting in a 98% service up-time between Q3 and Q4 2017, 6% higher than previous two quarters.
Education verified_user 0% verified
  • Confiz
    MLOps Fundamentals & LLM
    Confiz
    Jun 2023
  • Monash University
    Master of Business
    Monash University
    Jan 2018 - Dec 2019 (2 years)
  • L
    BS Electrical Engineering
    LUMS
    Aug 2011 - Jun 2015 (3 years 11 months)
Projects (professional or personal) verified_user 0% verified
  • Q
    Qualitative assessments for life science leaders using NLP
    Aug 2024 - Current (1 year 11 months)
    Microsoft Azure, Docker, ACR, Python, LLMs, Text extraction
  • Freelane
    Multi-Class Sentiment Classification for Online Marketing Agency
    Freelane
    Jul 2024 - Aug 2024 (2 months)
    Performed extensive EDA including clustering & PCA to understand the variance in data across different classes of sentiment on textual reviews with the labeled data.  Experimented with classical models including multinomial logistic regression, self-created neural networks with hidden layers using Tensor flow. Prepared modular ML Pipeline and deployed to Azure Machine Learning
  • C
    AI Chatbot & Recommendation System for Retail Store
    Confiz Limited
    Mar 2024 - Jun 2024 (4 months)
    Researched and developed AI Solution using Hugging face models & FAISS vector stores to provide product catalogue recommendations of relevant product from store catalogue for the retail customer.  Architect the AI solution using Azure Cloud services following RAG approach augmented with re ranker models.  Migrated and deployed the solution to Azure Cloud.  Designed AI based Chatbot for retail store’s corporate employees to support HR queries function with augmentation to Azure Open AI Service.  Deployed solution to Azure cloud for the retail store and integrate with existing web services application.
  • C
    AI based Information Retrieval for Large Construction Agency
    Confiz Limited
    Jan 2024 - Mar 2024 (3 months)
    Azure Document Intelligence, PaddleOCR, pdf2text, Python, OpenAI, DeepEval, LlamaIndex, Rule Mining, Python
  • M
    ML Forecasting & Anomaly Detection for Retail Giant
    Feb 2021 - Dec 2023 (2 years 11 months)
    Designed and developed a forecasting engine (including feature engineering, model training & tuning, model evaluation and predictions) to forecast long range forecast for channel-department sales and units for using statistical and Machine learning models including Prophet, SARIMAX & XgBoost.  Closely worked with Machine Learning Engineers to ensure smooth delivery to production following MLOps best practices at each cycle of the delivery.  Developed a supply chain forecasting module for replenishment of the FCs, and ACCs of the retail giant.  Developed and delivered a probabilistic anomaly scoring & detection module for sales & units data & forecasts.  Implemented data & model monitoring practices for the forecasting engine to producti