Work Type
Full-time employment
Location
Lebanon
Application Deadline
30 April
Apply
AI Engineer

 

We are seeking a talented AI Engineer to join our AI team and play a key role in building and shipping production-grade machine learning and Generative AI systems. This position focuses on two core pillars: machine learning, where you will train and ship models as a primary product surface, and agentic AI where you will build agent and RAG systems that power our GenAI products.

You will collaborate closely with a team of experienced engineers, actively contributing to architectural decisions while benefiting from continuous technical mentorship. The role requires end-to-end ownership of assigned components, encompassing data preparation, model development and evaluation, production deployment, and iterative improvement based on online metrics.

Furthermore, you will work in close coordination with cross-functional engineering teams, progressively assuming greater responsibility and contributing to the scalability, reliability, and overall advancement of our AI systems.

 

Specifications Required: 

Education: Bachelor’s Degree in Software Engineering, Computer Science, Computer Programming or Software Development. 

Experience: 2 to 4 years of software or ML engineering experience. Demonstrated ownership of at least one ML component in production with measurable online or offline outcomes. Hands on experience shipping at least one GenAI feature in production.  

Languages: Proficient in Arabic, French, English  

 

 

 

Your responsibilities:

 

Machine learning 

  • Build and ship ML models end to end across problems such as classification, regression, and ranking, with strong feature engineering, cross validation discipline and evaluation. 
  • Develop recommender pipelines: candidate generation (matrix factorization, two tower retrieval, ANN search) or ranking (learning to rank, gradient boosting, neural rankers). 
  • Drive quality improvements using offline metrics and online A/B testing. Handle practical challenges: cold start, popularity bias, feature freshness, and offline versus online metric drift. 
  • Pick the right tool for the job across gradient boosting, linear models, and, where it pays off, deep learning or LLM based approaches. 

 

GenAI and agents 

  • Contribute to our agents and RAG systems using LangGraph, Microsoft Agent Framework, or other agentic tools. Create skills, design prompts, and evaluate harnesses for specialized agents. 
  • Build RAG pipelines with hybrid search, chunking strategies, and rerankers on vector databases such as pgvector, Qdrant, or Pinecone. 
  • Apply structured outputs and tool calling to make LLM applications reliable. 

 

Data, MLOps, and LLMOps 

  • Build and maintain data pipelines that turn structured and unstructured data into training and serving features. 
  • Ship production services using Docker, CI/CD, and experiment tracking (MLflow or Weights and Biases). Follow our deployment patterns (canary, shadow, feature flags). 
  • Contribute to evaluation pipelines: help build golden datasets, LLM as judge rubrics, and CI integrated eval gates using tools such as Braintrust, Langfuse, LangSmith, or RAGAS. 
  • Use LLM observability tooling to debug, monitor, and improve systems in production. 
Qualifications you'll need:

 

A. Required Skills/Abilities: 

 

Core ML and languages 

  • Strong Python. Proficient SQL. 
  • ML fundamentals: training and validation splits, bias and variance, experimental design, and model evaluation.  
  • Deep learning: working knowledge with scikit-learn and pandas, plus Hugging Face Transformers for LLM use. 

– ML at depth: production experience with gradient boosting including cross validation strategy, feature engineering, leakage detection, calibration, and handling imbalanced data. 

– Recommender systems: production experience with either candidate generation (matrix factorization, two tower retrieval, ANN indexes such as FAISS, ScaNN, or HNSW) or ranking (learning to rank, gradient boosting, neural rankers). 

– Time series forecasting: working knowledge of methods such as ARIMA, ETS, Prophet, gradient boosting with lag features, or deep learning approaches such as Temporal Fusion Transformer. 

 

 GenAI basics 

– Hands on experience with at least one agent or RAG framework (LangChain, LangGraph, Google ADK, NVIDIA Agent Toolkit, Microsoft Agent Framework, or equivalent). 

– Working knowledge of at least one major LLM provider (Anthropic, OpenAI, or Google) and of structured outputs and tool calling. 

– Working knowledge of at least one vector database (pgvector, Weaviate, Qdrant, Pinecone, or Milvus). 

 

B. Preferred Skills/Abilities: 

  • Familiarity with post training techniques (PEFT, LoRA, QLoRA, or DPO) and willingness to go deeper. 
  • Inference optimization exposure (vLLM, SGLang, quantization, prompt caching). 
  • Exposure to guardrails and safety tooling (Llama Guard, NeMo Guardrails, Bedrock Guardrails). 
  • Basic AI governance familiarity (ISO/IEC 42001, EU AI Act at high level). 

 

C. Soft Skills/Abilities: 

  • Curious and collaborative. Asks questions, gives and receives feedback, and learns from the engineers around them. 
  • Clear communication written and verbal. Can explain tradeoffs to product and non technical partners. 
  • Reliable ownership of your workstream, with good judgment about when to escalate and when to push through. 

 

Work Conditions:  

Working hours: As per employment agreement. 

 

If interested, kindly share your updated cv to hr@inmind.ai mentioning name and position in the subject.