Franck Romuald MTOPE FOTSO

1. Developed robust ML pipelines for bespoke in-house models, enhancing automation and predictive accuracy [Google Cloud]. 2. Built multivariate demand forecasting models (VARIMA, TFT) for accurate supply chain predictions. Time series forecasting [Dart]. 3. Applied deep RL and multi-agent systems to enhance procurement and inventory management [MORL, MARL]. 4. Optimised logistics for a fleet of 25 vehicles handling 500 daily orders using RL and OR-Tools; achieved optimal route planning under constraints like delivery times, traffic conditions, and fuel consumption. 5. Customer & product trend analysis to drive upselling, personalised product recommendations, and targeted marketing campaigns. 6. Implemented Generative AI solutions leveraging GPT & Gemini to enhance customer engagement and product recommendation systems 7. Collaborated with business analysts and senior stakeholders to effectively integrate AI solutions with business operations. 8. Led and managed a team of 5 developers and 2 data scientists.

Conids LTD is an IT outsourcing company specializing in software development and consulting. Based in Yaoundé, Cameroon with a development center in Eastern Europe, we are committed to delivering advanced solutions to our clients. Conids LTD was established in 2019 and has quickly grown into a mid-sized company with extensive experience across various industries and technologies. Our team of highly qualified developers is equipped to handle even the most complex and specialized projects with ease. Our services: Visual design, Software Development, QA Testing, Digital Marketing, Dedicated development, Artificial Intelligence.

1. Supply chain mapping and identification of key digital enablers. 2. Design and build a framework (MAKE4ML) for eliciting knowledge from domain experts. 3. Set up infrastructure to support an ERP system within the business. 4. Establish a cloud-based data lake to enable the fusion of company data from diverse sources. 5. Tailor and integrate Transformer-based Time Series models for short & long-term stock forecasting. 6. Design, build, and train smart agents capable of making procurement decisions, such as replenishment of stock, choice of suppliers, and negotiated prices, using deep reinforcement learning algorithms like PPO, DDPG, and SAC. 7. Lead and manage a team of 5 developers, including full-stack, junior, and senior levels. 8. Publish journal & conference papers as outcomes of the KTP project.

1. Implementation of scientific / research software features in the medical platform DYNAMIKA 2. Medical image processing algorithm development 3. Prototype of new features and user-experiences 4. Work closely with clinical and academic partners and, when required, present the system at scientific conferences and customer meetings 5. Participate and provide progress reports on regular basis to assist in planning and monitoring of development 6. Design and Implement Machine Learning / Deep Learning pipelines improving decision making in clinical trials 7. Assist in other tasks associated with software engineering, software testing, and DevOps.

1. Tracking activities in construction sites using 360 cameras: designed and built tools for analysis and interpretation of equirectangular (panoramic) images. It allowed a fast and effective generation of reports for high rise residential projects in central London. [Python & TensorFlow] 2. Designed and implemented an instance-aware hashing tool to identify, quantify and store weekly changes (ceiling, floor, wall, etc.) into embeddings / binary codes. These embeddings were used to quickly analyze a new batch of images (collected weekly) and come up with similarity metrics. Techniques used: similarity metric learning, segmentation, the fusion of features, and truncated / adaptative backpropagation. [Python & TensorFlow]

Thesis: Designed and built a GIS-based software for the prediction and 3D visualization of mobile network coverage in a city (Case of Study: Yaoundé) [Java SE + Python + ArcGIS]