Applied AI Scientist - AVP

To be successful as an Applied AI Scientist, we would expect that you have: Required Qualifications: Any educational background from IIT will be preferred.Proven track record of working in AI, specifically in domains such as research, scientific publications and finance5+ years of experience with Python as a programming and scripting language 3+ years of experience with Machine Learning, Artificial Intelligence and Generative AIWorking knowledge and strong understanding of transformers, reasoning-focused model architectures, and inference-time compute strategies such as ReAct, chain-of-thought, tree-of-thought, self-consistency, Monte Carlo Tree Search (MCTS) for reasoningHands on experience with LLM fine-tuning, including reinforcement learning-based methods and preference alignment techniques (such as DPO, PPO, GRPO) and parameter-efficient fine-tuning methods (LoRA, QLoRA, adapters)Proven experience building and deploying agentic AI harnesses with multi-step reasoning, intent/model routing, tool-use, function calling, memory, etc.Proven experience taking use cases from ideation to production in the AI space at scale (Enterprise AIOps experience)Excellent stakeholder management and communication skills, as the person will shape the future interacting with the business and technology Desirable skillsets that would make you awesome: Model Training: Experience with PyTorch and Compilers, as we aspire to build new inhouse models, including distributed training, and familiarity with HuggingFace transformers, TRL and PEFT librariesAdvanced Training Paradigms: Experience designing and evaluating reinforcement learning-based fine-tuning pipelines for language models, including reward modelling, policy optimisation, and offline/online evaluation in complex domains such as finance Agentic AI and Orchestration: Experience with MCP, A2A and tool-use protocols, and orchestration frameworks such as LangGraph, CrewAI, ADK, or similarInference optimization: Knowledge of quantization frameworks (such as TurboQuant), serving frameworks (vLLM, SGLang, Triton), and latency/cost tradeoffsKnowledge Graphs: Experience designing and integrating knowledge graphs (Neo4j, TigerGraph, memGraph), GraphRAG, entity resolution, relationship extraction, and semantic retrieval to ground AI systems in enterprise contextVision Language Models: Experience working with and fine-tuning multimodal AI models, including image-document understanding, entity extraction, visual question answering, OCR-enhanced retrievalExperience in AWS and Bedrock (Generative AI), for agentic systems architectureExperience managing multiple projects at the same time, PMP Additional Skills that would be ideal: Ensure that all activities and duties are carried out in full compliance with regulatory requirements, Enterprise Wide Risk Management Framework and internal Barclays Policies and Policy Standards.Independent, reliable, motivated.Willing to learn and share knowledge.Experience with database systems, both relational and non-relational.Experience in the full development lifecycle, including design, implementation, testing, deployment, and support.Experience with version control systems like Git is desirable.Basic knowledge of Docker and Kubernetes.Experience with continuous integration systems like Jenkins or TeamCity. To be successful as an Applied AI Scientist, we would expect that you have: Required Qualifications: Any educational background from IIT will be preferred.Proven track record of working in AI, specifically in domains such as research, scientific publications and finance5+ years of experience with Python as a programming and scripting language 3+ years of experience with Machine Learning, Artificial Intelligence and Generative AIWorking knowledge and strong understanding of transformers, reasoning-focused model architectures, and inference-time compute strategies such as ReAct, chain-of-thought, tree-of-thought, self-consistency, Monte Carlo Tree Search (MCTS) for reasoningHands on experience with LLM fine-tuning, including reinforcement learning-based methods and preference alignment techniques (such as DPO, PPO, GRPO) and parameter-efficient fine-tuning methods (LoRA, QLoRA, adapters)Proven experience building and deploying agentic AI harnesses with multi-step reasoning, intent/model routing, tool-use, function calling, memory, etc.Proven experience taking use cases from ideation to production in the AI space at scale (Enterprise AIOps experience)Excellent stakeholder management and communication skills, as the person will shape the future interacting with the business and technology Desirable skillsets that would make you awesome: Model Training: Experience with PyTorch and Compilers, as we aspire to build new inhouse models, including distributed training, and familiarity with HuggingFace transformers, TRL and PEFT librariesAdvanced Training Paradigms: Experience designing and evaluating reinforcement learning-based fine-tu