AI​/ML Specialists; Data Scientists​/ML Engineer

Job Description

VAM Systems is currently looking for AI/ML Specialists (Data Scientists/ML Engineer) (On-Site) for our Bahrain operations with the following skillsets and terms & conditions:

7 – 10 years

Bachelor’s Degree in Computer Science / Engineering (Preferably BE Computer Science & Engineering)

• Machine Learning, Deep Learning, MLOps, AI in Financial Services.

Google Professional ML Engineer, Microsoft AI Engineer Associate Professional Licenses Required Not applicable.

• Tensor Flow Developer Certificate
• AWS Certified Machine Learning

• Proven hands‑on delivery experience in banking, financial institutions, or insurance within Gen AI solutions such as chatbots, document analysis, etc., leveraging RAG and robust architecture with proper governance and security measures
• Several years of ML experience with implemented use cases.
• Hands‑on work experience most of which in banking, financial institutions, or insurance industries.

• Ability to build and deploy ML models using Python and relevant libraries. Understanding of supervised and unsupervised learning algorithms.
• Experience with model evaluation and performance metrics.
• Familiarity with AI use cases in banking (e.g., fraud detection, personalization) Knowledge of data preprocessing and feature engineering.
• Ability to work with cloud‑based ML platforms (e.g., Azure ML, AWS Sage Maker). Understanding of MLOps and model lifecycle management.
• Ability to communicate insights and build explainable AI models.

Design and develop machine learning models to support AI‑driven banking solutions. Collaborate with data engineers to access and prepare data for modeling. Apply statistical and ML techniques to solve business problems (e.g., churn prediction, credit scoring). Evaluate model performance and optimize for accuracy, precision, and recall. Deploy models into production using MLOps frameworks and CI/CD pipelines. Ensure models are explainable, auditable, and compliant with regulatory standards.

Work with business stakeholders to identify AI opportunities and define success metrics. Document model assumptions, data sources, and performance benchmarks.

• Python (PyTorch, Tensor Flow, Lang Chain, Hugging Face, OpenAI API, Anthropic Claude, etc.)
• LLM fine‑tuning (LoRA, PEFT, prompt tuning)
• Retrieval‑Augmented Generation (RAG), vector databases (Pinecone, FAISS, Weaviate, Chroma)
• Prompt engineering and orchestration (Lang Chain, Llama Index, Semantic Kernel, DSPy)
• Knowledge of embeddings, tokenization, and transformer architecture
• Cloud AI tools: AWS Bedrock, Azure OpenAI, Vertex AI, Open Search, Elastic Search
• Model evaluation: hallucination detection, grounding, and benchmarking (BLEU, ROUGE, Truthful
  
  QA, etc.)

• RESTful and Graph
  
  QL APIs, webhooks
• Containerization and deployment (Docker, Kubernetes, CI/CD)
• Authentication and user/session management
• Data pipelines and microservices
• Knowledge of frameworks like FastAPI, Flask, NestJS, or Express
• Integration with enterprise data (SharePoint, Salesforce, SQL, internal APIs)

• Lang Graph, Auto Gen, CrewAI, Flowise, or similar agent frameworks
• Task‑decomposition and reasoning chains
• Function calling, tool use, and API chaining
• Memory design (short‑term vs long‑term)
• Context management and grounding strategies.

• Conversation design frameworks (Google CCAI, Microsoft Bot Framework, Voiceflow, Botpress)
• Flow design and intent management (Dialogflow, Rasa, Cognigy)
• Tone, empathy, and personality design for AI personas
• A/B testing dialogue variants and measuring user satisfaction.

• Data pipelines (Airflow, dbt, Kafka)
• Structured/unstructured data ingestion (PDFs, databases, APIs)
• Feature store and model registry management (MLflow, Kubeflow)
• Vector database deployment and optimization
• Monitoring, logging, and drift detection.

• Model explainability (SHAP, LIME)
• Bias/fairness audits and data privacy
• Compliance with GDPR, ISO 42001, NIST AI RMF, and local banking regulations
• Secure prompt logging and audit trails.

• Translating business problems into AI use cases
• Road mapping and budget planning
• KPI design (accuracy, user satisfaction, automation ROI)
• Vendor management (OpenAI, Anthropic, AWS, etc.)
• Change management and user adoption

15 - 30 days