تخطي إلى المحتوى الرئيسي
Maryam Mohamed

Maryam Mohamed

Junior AI Engineer | LLM & RAG Systems | Computer Vision | FastAPI

كفاءة
Cairoعضو منذ ١١ أكتوبر ٢٠٢٦
PythonSQLPostgreSQLMySQLC++LangChainFAISSSentence-TransformersHugging FaceOllamaembeddingssemantic searchreranking (bge-reranker)multi-agent orchestrationPyTorchTensorFlowTransformersBERTCNNRNN/LSTMOpenCVYOLOSAM segmentationOCR (Tesseract)PDF/image processingdocument extractionSupervised/unsupervised learningtime-series forecasting (LSTM, DeepVAR)feature engineeringhyperparameter tuning
·

وظّف هذا الشخص

أرسل عرض عمل مباشرة لهذا المرشح

PythonSQLPostgreSQLMySQLC++

نبذة

Junior AI Engineer with a B.Sc. in Artificial Intelligence and hands-on experience building LLM/RAG applications, computer vision systems, and FastAPI microservices. Delivered a chick-counting vision system at 95% count accuracy, a multilingual RAG document-intelligence platform, and a National ID OCR service integrated into a live application. Strong in Python, PyTorch, LangChain, FAISS, Docker, and PostgreSQL.

الخبرة

Junior Data Scientist at Datalentech (2026-02 – Present)

  • Built an automated chick-counting computer vision system on the Count-anything model, achieving 95% count accuracy on production imagery and replacing manual counting; reduced duplicate detections on dense, overlapping chicks by tuning inference and post-processing parameters.
  • Developed the Cost Service as a FastAPI microservice with REST endpoints, business logic, data validation, and integration with the platform’s other services.
  • Automated slaughterhouse pricing and profitability decisions through cost-factor, margin, and profitability calculations, plus optimization logic balancing feed allocation and product margins.
  • Designed the PostgreSQL data layer with SQLAlchemy (schemas, data models, queries, persistence workflows) for cost and pricing data.

AI Engineer Intern at ADVEC (2025-07 – 2025-08)

  • Built a module in a multilingual platform that lets users summarize, query, and generate questions from PDF documents.
  • Engineered an end-to-end RAG pipeline (LangChain, FAISS, MiniLM embeddings, LLMs) with document chunking, embedding generation, vector storage, and similarity-based retrieval for conversational Q&A.
  • Developed summarization and question-generation pipelines supporting multiple difficulty levels.

Data Scientist at Intro Resource Recovery (2024-07 – 2025-03)

  • Built an end-to-end National ID OCR and verification system (OpenCV, Tesseract, FastAPI, Docker) that automated document processing and reduced manual data entry; integrated into the DAWAR application.
  • Developed and optimized ML forecasting models for broiler weight, feed consumption, and mortality, improving prediction accuracy through feature engineering and iterative evaluation.
  • Designed end-to-end data pipelines (preprocessing, EDA, feature engineering, validation) producing production-ready datasets.
  • Built and integrated RESTful APIs exposing data and ML capabilities to downstream applications.

التعليم

B.Sc. in Artificial Intelligence – Nile University (2026)

المهارات

Junior AI EngineerLLM & RAG SystemsComputer VisionFastAPI

التقييمات

ذات صلة