• Research and prototype innovative machine learning approaches in areas such as Search, Large Language Models (LLMs), Multimodal Content Understanding, and Recommender Systems • Design and implement end-to-end ML pipelines—from data exploration to model training, evaluation, and deployment • Analyze large-scale datasets to identify opportunities for personalization and improved user experiences • Partner with Product, Design, and Engineering teams to integrate models into Dropbox products • Contribute to the team’s technical discussions, offering research-based perspectives to guide experimentation and long-term strategy

• Architect and implement sophisticated multi-agent systems and autonomous workflows leveraging the Google AI SDK, LangGraph, and LangChain to solve complex, non-linear business processes. • Lead the design and construction of cloud-native solutions, using Terraform, Kubernetes, and Docker to ensure that AI models are deployed on scalable, reliable infrastructure. • Apply rigorous statistical evaluation frameworks to model performance, moving beyond standard metrics to include uncertainty estimation, calibration, and robust hypothesis testing during model optimization (LoRA, QLoRA). • Lead the development of custom predictive models and deep learning solutions, utilizing frameworks like PyTorch and Scikit-Learn to select suitable architectures—whether decision trees, neural nets, or ensembles—based on performance and client criteria. • Design and implement state-of-the-art generative models for NLP and multimodal tasks, leveraging tools like OpenCV for image preprocessing and Stable Diffusion concepts where applicable. • Champion MLOps best practices within the team, building validated data pipelines and CI/CD/CT workflows using Kubeflow and Vertex AI Pipelines to ensure model quality and integrity. • Collaborate directly with clients to understand their unique needs, translating business challenges into technical solutions and providing expert guidance on dataset management best practices. • Personally tackle the most difficult engineering challenges, identifying technical risks such as overfitting or latency issues, and optimizing hyperparameters to ensure precision and interpretability.

• Architect and lead end-to-end computer vision projects focused on: Equipment defect detection, Thermal anomaly identification, Vegetation encroachment monitoring, Surveillance of closed areas for human and animal intrusions • Drive innovation by incorporating the latest advances in deep learning and generative AI to enhance model training, accuracy and reliability • Develop production-grade Python libraries for the complete ML lifecycle • Mentor team members and establish best practices for model development, evaluation, deployment, and monitoring • Advocate for and uphold software quality standards within the ML team

• Design and implement architecture for multi-step planning and execution • Own domain-specialised training and evaluation infrastructure • Build quality gates for performance monitoring and deploys