Mridul Ahluwalia

ML Optimization: Working in a consumer DNA project for a US-based food and beverage company, I optimized the ML pipeline, reducing operational time by 90%. This included updating preprocessing logic, implementing faster training models, decoupling data-agnostic processing, and refining Spark logic. I elevated coding standards, updated the CI/CD pipeline, wrote comprehensive unit tests, and managed tasks through Jira. To enhance robustness, I implemented model drift tracking and introduced MLflow for experiment management. I improved the logging mechanism and applied performance optimizations, significantly enhancing the ML infrastructure.

Productionize ML Pipelines: As an SME, I contributed to multiple projects, including leading a team to construct an ML pipeline for a UK-based bank, focusing on AWS SageMaker for model training, inference, and drift monitoring.

Platform Creation: I contributed to establishing a unified AWS cloud-based ML platform for a US-based healthcare company, executing various POCs including Feature Store setup and Dataiku ML pipeline implementation.

Mentorship & Technical Presentations: I mentored new trainees, juniors and team members, delivered technical presentations to clients and management.

As a Python team lead and developer specializing in ML and IoT applications, I delivered innovative solutions and led key projects. I spearheaded a warehouse automation initiative using Python, Django, AWS, and IoT, enhancing operational efficiency. I engineered an NLP model for resume parsing and a bespoke named-entity-recognition system with Spacy and Gensim. I established an adaptable automated pipeline for ongoing model refinement and led the development of a sophisticated visitor management system for residential communities. Additionally, I designed advanced IoT modules for water dispenser ATMs, created a robust backend architecture with Python, C, and Raspberry Pi, and orchestrated the development of a smart IoT device management platform using AWS services to establish a serverless backend infrastructure.

Ensured smooth operation of trading systems, optimized infrastructure, and resolved data accuracy issues using in house tools and SQL