Charlin DJIOKO

6-month end-of-studies internship within the Model Design team at RISK BCEF Architecture, responsible for the development and monitoring of credit risk models across mortgage, consumer credit and SME portfolios. Key contributions: • Rebuilt the SME origination scoring model: more parsimonious architecture currently being pushed to production. • Improved the BNP's internal Python package for PD, LGD and CCF modelling: memory optimisation, performance gains, and training of the Model Design team. • Proposed decision thresholds and translated statistical performance into concrete operational impacts. • Contributed to backtesting of models in production and ensured full traceability of all work produced.

Developed an AI & cloud-based solution to enhance the differentiation of corporate default risk through large-scale transactional data enrichment. Key contributions: • Designed a full NLP pipeline: preprocessing, embeddings (SBERT), clustering (UMAP, HDBSCAN), zero-shot LLM classification and semantic search (FAISS) — deployed on Google Cloud (BigQuery, Cloud Run, Artifact Registry). • Architecture strongly limiting exposure to external LLMs, with the large majority of transactions processed through semantic search. • Improved default risk model performance through fine-grained transactional categorisation. • Reduced a process initially estimated at several years to just a few hours at marginal cost.

I prepare and run courses for high school and college students, identifying their learning difficulties and implementing effective remedial strategies. I help them to succeed.

Developed a dynamic predictive model to estimate relapse risk in patients with PTTi, based on real-world longitudinal hospital data. Designed a personalized risk scoring system, updated in real time, to support clinical decision-making. Applied advanced statistical methods, including Bayesian updating and individual patient trajectory modeling, using R and Python.

"Young people's knowledge and use of AI." Oversaw data collection for the annual ISSEA survey, coordinating teams to gather data from over 1978 households. Played a key role in questionnaire design, data collection app development, and automating the training process for data collection teams using Python.

Designing and facilitating mathematics, physics, chemistry, and computer science courses; assessing and addressing student learning gaps.

Automated the monthly computation of inflation indicators using Excel VBA, improving efficiency and reducing calculation errors. Processed and cleaned national consumer price index (CPI) data for statistical reporting. Produced dashboards and contributed to an analytical report on inflation trends in Cameroon. Collaborated with economists and statisticians to ensure data quality and relevance for macroeconomic monitoring.

I prepare Cameroonian students for entrance exams to top universities and schools