Moazam Ashraf

• Increased customer satisfaction rates through diligent collaboration with cross-functional teams on feature development and bug resolution.
• Mentored junior engineers to strengthen their technical skills and improve overall team efficiency.
• Completed code debugging, software troubleshooting, root cause analysis, and program reviews.
• Unified efforts of disparate engineering and design teams by applying collaborative software review processes.
• Participated in architectural discussions to build confidence and promote customer success when developing new and migrating existing applications, software and services on AWS platform.
• Streamlined software development processes by integrating agile methodologies and automating repetitive tasks.
• Developed customized AI algorithms tailored to specific client requirements, resulting in more effective solutions and higher customer satisfaction.

Generative AI

Developed 3 different end to end RAG solutions with multilingual capabilities and multiple industry and domain verticals (Arabic , Turkish), Utilizing Anthropic Claude , Jais , Zephyr , Bloomz.

Finetuned 7 billion , 13 billion , 40 billion parameters based LLMS for domain adaption , instruction tuning using Sagemaker Finetuning I and then Lora(low ranking adaption) to Tune LLM's for Project needs.

Serving LLM's and Embedding Models on AWS Sagemaker real time inference endpoints, also customizing the deployment containers based on deepseed/mpi in case of non availability of containers in Selected Regions of AWS.

Computer Vision

Developed the ML backend for a POC involving aircraft management , using YOLO architecture to identify and label all events taking place on the runway of an airport and all the actions performed for maintenance of the aircraft.

Working with a large-scale public entity pertaining to law and order leading the Research and Development team and dealing with vendors and research partnerships with other public entities and universities. Lead Team and Deployed indigenous Automated number plate recognition, Road Congestion Using YOLO architecture and OpenCV and Python as the main platform, team building and management followed by working on Machine learning and Natural Language Processing was the best bit about the learning curve

• Collaborated with cross-functional teams to successfully deliver comprehensive solutions for clients.
• Managed client relationships through regular check-ins and updates on project progress.
• Evaluated client needs and expectations, establishing clear goals for each consulting engagement.
• Developed innovative strategies for clients, resulting in increased revenue and business growth.
• Having good team coherence and communication to translate requirements and optimizing the solutions.

Working with this Multiple Departments to deliver software solutions and reporting to all stakeholders while working with the dev and QA team, the first foray into Managerial aspect on my end and exposure to Agile Methodologies and Project Management.

• Sprint Planning and Scrum Exposure.

• Using Pandas to optimize existing Data Inflow and outflow.

• Replacing trivial Tasks with Python Scripts to instill basic automation.

• Working with large scale data pertaining to agriculture and Citizen Profiling.

• Coordination with client and then the whole Dev team to have a well refined SRS documentation.

• Having Technical Background participating in the development work and optimizing the devised solutions and plans.