Senior AI Engineer

The core responsibilities for the job include the following:

Generative AI and Agent Development:

• Design and build LLM-powered applications and AI agents.
• Develop workflows using LangChain / LangGraph / LlamaIndex.
• Implement: Function calling/tool usage, chain-of-thought/reAct reasoning.
• Build end-to-end agentic systems capable of multi-step reasoning

.

RAG (Retrieval-Augmented Generation) Systems:

• Design and implement scalable RAG pipelines.
• Work with: Vector databases (Pinecone, FAISS, Milvus, etc. ).
• Handle: Document chunking strategies, embedding generation, and retrieval optimization.
• Improve response quality and reduce hallucinations.

LLM Optimization and Fine-Tuning:

• Fine-tune models using LoRA/PEFT techniques.
• Optimize performance via quantization (8-bit / 4-bit).
• Manage context window limitations and prompt engineering strategies.
• Work with alignment techniques like RLHF (understanding level expected).

Deep Learning and Model Understanding:

• Strong grasp of neural networks, backpropagation, and regularization techniques (dropout).
• Experience with LSTMs/RNNs (foundational understanding) and CNNs (for vision-based use cases, if applicable).
• Train and optimize models using GPU/TPU environments.

Machine Learning and MLOps:

• Apply ML techniques for structured and unstructured data.
• Strong understanding of class imbalance problems and evaluation metrics (recall, precision, etc. ).
• Work with XGBoost / Random Forest for tabular problems.
• Handle model monitoring and data drift/concept drift detection.
• Build production-ready ML pipelines.

Requirements:

• 2-5 years of experience in AI / ML / GenAI engineering.
• Strong hands-on experience with Python and LLM frameworks (LangChain, LlamaIndex, etc. ).
• Experience building: RAG pipelines, AI agents, and tool-using systems.
• Understanding of: Vector databases, prompt engineering, and hallucination mitigation.
• Solid ML/DL fundamentals.

Good to Have:

• Experience with fine-tuning LLMs (LoRA/PEFT), quantization and optimization techniques, and GPU-based model training.
• Exposure to MLOps tools and deployment pipelines, and real-time AI applications.