Lead Data Scientist

Job Responsibilities: 

• Collaborate with product and business teams to identify problems and define scalable, measurable AI/ML-driven solutions. 
• Evaluate, fine-tune, and deploy production-ready models using the latest open-source and commercial model families (e.g., GPT-4o, Claude 3, Mixtral, LLaMA 3, Gemini). 
• Develop and manage domain-specific generative models for tasks like summarization, classification, extraction, and generation using Transformers and LLMs. 
• Build and maintain retrieval-augmented generation (RAG) pipelines, including vector databases (e.g., Weaviate, FAISS, LanceDB, Pinecone). 
• Work with frameworks such as LangChain, LlamaIndex for prompt chaining, agent orchestration, and memory-enabled AI. 
• Optimize for commercial KPIs such as inference cost, latency, model accuracy, and scalability. 
• Benchmark and profile models to ensure high performance and reliability in production environments. 
• Keep up to date with the evolving landscape of AI/ML and recommend tools, models, or practices that can improve team capabilities. 
• Understand business requirements to propose and develop scalable and effective commercial solutions. 
• Translate stakeholder needs into actionable data science roadmaps, prioritize deliverables, and plan resources across multiple initiatives. 
• Assist, mentor, and train junior and mid-level data scientists to foster skill development, code quality, and cross-functional collaboration. 

Skills & Competencies Required: 

• Programming: Expert-level proficiency in Python and common ML/AI libraries (NumPy, Pandas, scikit-learn, PyTorch, TensorFlow, Hugging Face Transformers, LangChain). 
• LLMs & NLP: Deep experience working with transformer-based architectures and LLMs, including prompt engineering, fine-tuning, and instruction-tuning. 
• Generative AI: Practical knowledge in building GenAI applications using both open-source (Mistral, LLaMA) and API-based (OpenAI, Anthropic) models. 
• RAG & Vector Search: Experience with RAG architecture and vector databases like FAISS, Pinecone, or Qdrant. 
• Machine Learning & Deep Learning: Solid understanding of supervised, unsupervised, and reinforcement learning algorithms. 
• Ability to work independently with minimal supervision in a dynamic and deadline-sensitive environment