Ntt Data Aivista Member of Technical Staff at Ntt Data Aivista designing and deploying agentic AI systems for enterprise customers. Customize foundation models and build scalable software for regulated environments.
Responsibilities
Design and deploy agentic systems — build the agentic infrastructure and workflows that power our enterprise AI products.
Customize models to the domain — adapt foundation models to each customer through tuning, RAG, and related techniques.
Engineer for production — build scalable, reliable software systems that keep AI products running under real enterprise load.
Deploy alongside customers — partner with cross-functional and customer teams to ship capabilities into enterprise environments.
Iterate relentlessly — improve capabilities, troubleshoot technical issues, and turn customer feedback into product improvements.
Representative Projects
Ship an agent that runs a regulated back-office process end-to-end — pulling from the customer’s systems of record, applying their policies and risk classifications, and executing with a human approval step where the stakes demand it.
Build the last-mile specialization layer that teaches a foundation model a customer’s domain — their workflows, client classifications, and regulatory interpretations — and prove it holds up on their real cases.
Design the agent infrastructure — orchestration, tool integration, memory, and context — that lets agents act reliably across a customer’s real applications and data.
Stand up the evaluation and guardrail harness that decides whether an agent is reliable enough to run in production against live enterprise systems.
Instrument governance and observability so every agent decision is auditable, policy-aligned, and defensible to a regulator.
Drive down the latency and cost of agentic workflows running at enterprise scale without giving up reliability.
Qualification
Bachelor’s degree in a technical fieldExperience with model optimizationsExperience with LLM evaluationYou don’t need to check every box
Required
8+ years building and shipping AI products, including systems running in production
Bachelor’s degree in a technical field, or an equivalent combination of education, training, and experience.
Production-quality software engineering in Python and other modern languages (Go, TypeScript, Rust, or similar), with a track record of taking systems from zero to production — durable software, not one-off scripts
A track record of taking AI capabilities from prototype to production — agentic workflows, RAG, or model customization — that stand up to real enterprise use
Hands-on experience deploying and operating services on a major cloud platform (AWS, Azure, or GCP), using containers and orchestration (Docker, Kubernetes, Helm, or similar)
Hands-on experience building LLM-powered and agentic applications — model APIs (e.g., OpenAI, Anthropic, Amazon Bedrock), agent frameworks (e.g., LangGraph, LlamaIndex), and retrieval-augmented generation (RAG)
Bias for action and ownership — you thrive with high autonomy and ambiguity, and ship without sacrificing rigor
Clear communication skills — comfortable working directly with enterprise customers and across engineering, product, and research
Preferred
Hands-on design and implementation in an industry-leading AI product, as the technical leader for important product capabilities
Experience with model optimizations, agent protocols, and large-scale data processing
Experience with LLM evaluation, fine-tuning, or applied NLP
Experience deploying AI or software in large, regulated enterprise environments
Advanced degree (MS or PhD) in Computer Science, AI, or a related field
A track record of independent work that demonstrates AI depth — shipped products, open-source contributions, research, or technical writing
You don’t need to check every box. If you’re excited about this work and confident you can do it — even if your experience doesn’t line up with every qualification above — we’d rather hear from you than have you rule yourself out. Strong candidates often bring backgrounds we didn’t expect.