Farhan Jawaid

UKDIRECT:

• Designed and crafted multiple microservices dedicated to OCR mappings for enhanced ID verification documents including Residence Permit, Driving License and Passport.
• Pioneered the implementation of a versatile algorithm capable of extracting information from the Machine Readable Zone (MRZ) Code on passports from various countries.

PKDIRECT:

• Implemented open source PaddleOCR on PAK CNIC SMART to replace paid OCR services e.g. Google & AWS OCR Services. Primary goal is to cut the Costing Plan

Miscellaneous:

• Engineered Code Obfuscation Technique for On Prem Deployments.
• Integrated yolov7 Model of Screen Detection into Torchserve's mar file.
• Configured Service Status and Error Alerts on Grafana That helped to take prompt action on any unexpected downtime and Unexpected Errors.
• Configured Daily API Endpoint Count Report and Ram Usage Report of Containers Running on UAT and PROD.
• Developed Conversation chatbot for POC level requirement using AWS Lex and Knowledgebase

• Facilitated integration of machine learning services into technology infrastructure.
• Gathered customer requirements and developed machine learning product features to address needs.
• Designed and deployed machine learning models ( supervised learning ) into product stack in support of product development.
• Partnered with Computer Vision Engineer to create model training workflows for use by end-users.
• Initiated A/B tests and regression testing to validate model performance using MLOps (Mlflow) prior to production for US and Local client.

• Analyzed business requirements into technical requirements for reporting and visualizations, data models and databases.
• Developed Flask API for Machine Learning use cases
• Developed company-wide technical strategy plans for Python-based assets, overseeing development, coding, debugging and deployments.
• Introduced cloud-based technologies into Python development to expand on-premise deployment options.
• Dockerization of Machine Learning use cases

• Data Scrapping of 4000 + doctors data from web based sources
• Adhoc Queries on SQL ( simple and complex queries )
• Daily Conversion Rate for E-Commerce Product
• Peak Hours Analysis
• Customer Churn rate ( 6 months )
• Extraction of GMV and NMVs Value

• Extraction and Transformation of EFPIA Compliance Health Care Professionals (HCPs) and Health Care Organizations ( HCOs) from PDFs.
• Data Scrapping of HCPs and HCOs from Web Source.
• Data Cleaning on Excel Sheet as per required SOPs.
• Adhoc queries on MSSQL for Quality Assurance.