Job Description
At EY, you’ll have the chance to build a career as unique as you are, with the global scale, support, inclusive culture and technology to become the best version of you. And we’re counting on your unique voice and perspective to help EY become even better, too. Join us and build an exceptional experience for yourself, and a better working world for all.
Job Description – AI Research Engineer, EY GDS India
Job title-AI Research Engineer
Experience-10+ years overall
About the role
We are looking for a highly motivated AI Research engineer to drive Innovation in AI, Generative AI and agentic AI. The role focuses on tracking and evaluating emerging technologies, building proofs of concept and working demonstrations, conducting pilots, and translating innovative ideas into enterprise-ready solutions and engineering patterns.
You will track what is moving in the AI ecosystem including but not limited to model adaptation, agentic AI, retrieval, reasoning, evaluation, AI safety, LLM Ops and AI-assisted engineering. You will judge what matters for enterprise clients, then prove it with code. Your solutions and POVs will shape EY's AI offerings, support pursuits, and give clients a clear view of what is real and how to scale emerging technologies.
This is an applied research and engineering role, not a foundational or academic research position. Success will be measured by the speed and quality with which emerging capabilities are converted into evidence, reusable assets, differentiated points of view and enterprise adoption paths.
Key responsibilities
Trend scanning and rapid evaluation
- Continuously track the AI landscape: frontier and open-weight model releases, research papers, open-source projects, developer tools, agent platforms etc
- Follow signals from research labs, conferences, developer communities, analyst reports and startup activity.
- Spot shifts early, such as new model capabilities, falling cost curves, emerging architectures and new ways of building software, and assess their impact on enterprise adoption.
- Maintain a living AI tech radar and share regular trend digests with practice leaders and delivery teams.
Hands-on experimentation across emerging AI areas
Build quick prototypes across emerging AI areas. The examples below are indicative, not exhaustive:
- Agentic AI and harness engineering: Design the AI harness around models, including prompts and context assembly, tool integration, memory and state management, planning and orchestration, structured outputs, guardrails, fallback logic, human-in-the-loop controls, and interoperability protocols such as MCP and A2A.
- Model adaptation, SLMs and efficient learning: Evaluate and adapt small and open-weight language models using fine-tuning, LoRA/QLoRA, instruction tuning, distillation, quantization, pruning and model merging for domain-specific, cost-efficient enterprise use cases.
- Reasoning, planning and verification: Experiment with reasoning models, test-time compute, planning, reflection, self-correction and verifier-based approaches to improve the reliability of complex analysis, decision-support and task-execution workflows.
- Evaluation, safety, assurance and governance engineering: Develop evaluation harnesses, benchmarks, golden datasets and automated quality gates covering accuracy, groundedness, hallucination, bias, robustness, safety and agent task completion, with traceability and governance evidence.
- AI-assisted software engineering and computer use: Experiment with coding agents, repository-level reasoning, specification-driven development, code review, testing, modernization
- LLMOps, inference optimization and AI economics: Evaluate model deployment, observability, continuous evaluation, routing, caching, batching and inference optimization, while assessing token consumption, latency, infrastructure cost and cost per business outcome.
- AI security, privacy and supply-chain assurance: Assess prompt injection, data leakage, insecure tool use, agent identity and permissions, model and dataset provenance, open-weight model risks, dependency vulnerabilities, confidential inference and privacy-preserving AI architectures.
- Emerging AI frontiers: Track and rapidly evaluate developments such as new model architectures, long-context and persistent-memory systems, compound AI, model routing, AI-native interfaces, neuro-symbolic approaches and evolving enterprise AI platforms.
- Retrieval, knowledge and enterprise context: Build and assess advanced RAG, Graph RAG, knowledge graphs, semantic search, reranking, long-context processing, enterprise memory and context-management approaches for grounded and domain-aware AI solutions.
Demos, POVs and accelerators
- Build end-to-end demos of enterprise use cases using emerging AI technologies.
- Write concise POVs, blogs and client briefings that explain a trend, its business relevance, risks and a recommended path.
- Package reusable assets (reference architectures, starter repos, eval sets) for delivery teams.
- Present to client stakeholders, EY partners and pursuit teams, support RFP responses with technical depth.
Responsible AI
- Build in explainability, audit trails, guardrails and data protection from the first prototype.
- Partner with Risk and Responsible AI teams on bias, fairness and data-handling reviews.
Required skills and experience
- Strong Python and software engineering skills
- Hands-on LLM application building with at least one agentic framework such as LangGraph, LlamaIndex, Semantic Kernel, AutoGen / AG2, CrewAI or a provider agent SDK.
- Practical fine-tuning of open models including data preparation and evaluation.
- Broad, current awareness across the emerging AI focus areas, with proven depth. Able to read a paper or release note, extract what matters and ship a working prototype quickly.
- Clear written and spoken communication, able to explain trade-offs to both engineers and business leaders.
Why Join Us
- Be at the forefront of AI-driven innovation across multiple client sectors.
- Work with global clients to drive real business impact.
- Collaborate with a team of AI experts, analytics leaders, and industry specialists in a highly entrepreneurial environment.
What we offer
EY Global Delivery Services (GDS) is a dynamic and truly global delivery network. We work across six locations – Argentina, China, India, the Philippines, Poland and the UK – and with teams from all EY service lines, geographies and sectors, playing a vital role in the delivery of the EY growth strategy. From accountants to coders to advisory consultants, we offer a wide variety of fulfilling career opportunities that span all business disciplines. In GDS, you will collaborate with EY teams on exciting projects and work with well-known brands from across the globe. We’ll introduce you to an ever-expanding ecosystem of people, learning, skills and insights that will stay with you throughout your career.
- Continuous learning: You’ll develop the mindset and skills to navigate whatever comes next.
- Success as defined by you: We’ll provide the tools and flexibility, so you can make a meaningful impact, your way.
- Transformative leadership: We’ll give you the insights, coaching and confidence to be the leader the world needs.
- Diverse and inclusive culture: You’ll be embraced for who you are and empowered to use your voice to help others find theirs.
EY | Building a better working world
EY exists to build a better working world, helping to create long-term value for clients, people and society and build trust in the capital markets.
Enabled by data and technology, diverse EY teams in over 150 countries provide trust through assurance and help clients grow, transform and operate.
Working across assurance, consulting, law, strategy, tax and transactions, EY teams ask better questions to find new answers for the complex issues facing our world today.
More Information
Application Details
- Organization DetailsErnst & Young


Recommended Comments
There are no comments to display.
Join the conversation
You are posting as a guest. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.