Description

Shape ZenseAI.QI and ZenseAI.AssureAI

- Architect and evolve the platform suite across both engines: the 18-agent ZenseAI.QI stack (DeepSpeci, CaseGeni, Auto-PlayPilot, DataGeni, Secure-Xi, Perf-Xi, Insights360, RIA and Accessibility Intelligence) and the four-pillar ZenseAI.AssureAI framework (Data Quality, Model Evaluation, Trustworthiness, and Scalability & Security).

- Design and refine the archetype-specific assurance lifecycles for Classical ML, Generative AI and Agentic AI, including the signature eight-axis agentic trajectory scorecard graded on every run.

- Keep the platforms LLM-agnostic and deployable on any client stack — on-prem, cloud or hybrid — as model-mixing and constant provider churn become the market norm.

- Own technical roadmap decisions for accelerators such as the Agentic Foundry (a 25-blueprint reference matrix) and the curated, swap-ready tooling ecosystem around each archetype.

Lead Evaluation-Driven Development

- Build and govern eval suites — ground-truth Q&A; sets, LLM-as-judge rubrics, frozen baselines — that gate every release rather than validate it after the fact.

- Own trajectory grading, red/purple/blue-team probes and safety attestations for agentic and generative systems, and drift monitoring and fairness audits for classical ML.

- Translate evaluation results into release decisions: eval-threshold gates, red-team severity floors, canary and shadow deployments, and rollback rehearsals.

- Bring evaluation-driven development practice into client engagements — showing, not just telling, how a live harness beats a one-time audit.

Own Presales and Client Proposals

- Respond to RFPs, RFIs and client proposals across both engines, translating client requirements into a defensible solution architecture and commercial structure.

- Architect engagements across the full ladder — AI QA Assessment, AI QA Transformation, Managed AI QA — and fast-starts such as the LLM Health Check, Agent Stress Test and Compliance Sprint.

- Build estimates, staffing plans and technical win themes that hold up under client and internal scrutiny, across client-managed, risk-reward and Zensar-managed commercial models.

Present and Articulate Value to Clients





- Present the ZenseAI.QI and ZenseAI.AssureAI value proposition directly to client stakeholders — from QA Directors and Heads of AI/ML Engineering to CTOs, Chief Risk Officers and Chief AI Officers — adapting the narrative to each audience.

- Lead client workshops and technical walkthroughs, including guiding a CIO through the Agentic Foundry in a single session.

- Build and deliver executive trust scorecards and portfolio risk heat maps that make the assurance story board-ready.

Represent Zensar in Client Captives for Sales

- Present live in client captive sessions and sales pursuits as the senior technical voice for Quality Intelligence, alongside account and sales leadership.

- Run and narrate proof-of-concept demonstrations that turn a client's stated risk into a small, credible proof — the practice's proof-first approach to closing.

- Represent the practice at partner and industry forums as the practice's thought leadership and pipeline develop.

Build Practice IP and Mentor the Next Generation

- Contribute reusable accelerators, reference architectures and industry packs across BFSI, TMT, and Manufacturing & Retail back into the practice's IP base.

- Mentor and help build out the practice's emerging AI-specialist roles — Prompt Engineer, LLM-Eval Engineer, Agent Architect, Trajectory Eval Engineer, Knowledge/RAG Engineer, AI Security Analyst and Adversarial Red/Blue Team Lead.

- Support GenAI and AI-enablement training for QI associates across the practice.

Responsibilities

Must-Have

- Approximately 17–21 years of overall technology experience, including a substantial, hands-on foundation in Quality Engineering / Test Engineering — automation architecture, test strategy and the full software testing lifecycle — not AI experimentation layered on top of a thin QE background.

- Working architectural knowledge of agentic AI systems: multi-agent orchestration, tool use, memory,



and the operational risks specific to agents — unauthorised tool calls, trajectory drift, memory leakage across accounts.

- Direct experience with evaluation-driven development: building or governing eval suites, LLM-as-judge frameworks, ground-truth datasets, frozen baselines and red-teaming / adversarial testing for generative or agentic systems.

- Strong client-interfacing capability: comfortable owning technical content in an RFP response, presenting to CXO-level stakeholders, and holding the room in a live client or sales captive setting.

- Ability to talk convincingly across both deterministic QE and probabilistic AI assurance — the “one estate” narrative is the core sell, and the architect has to embody it, not just recite it.

- Familiarity with the regulatory and standards backdrop shaping AI assurance conversations: EU AI Act, NIST AI RMF and ISO/IEC 42001.

Qualifications

- Approximately 17–21 years of overall technology experience, including a substantial, hands-on foundation in Quality Engineering / Test Engineering — automation architecture, test strategy and the full software testing lifecycle — not AI experimentation layered on top of a thin QE background.

- Working architectural knowledge of agentic AI systems: multi-agent orchestration, tool use, memory, and the operational risks specific to agents — unauthorised tool calls, trajectory drift, memory leakage across accounts.

- Direct experience with evaluation-driven development: building or governing eval suites, LLM-as-judge frameworks, ground-truth datasets, frozen baselines and red-teaming / adversarial testing for generative or agentic systems.

- Robust client-interfacing capability: comfortable owning technical content in an RFP response, presenting to CXO-level stakeholders, and holding the room in a live client or sales captive setting.

- Ability to talk convincingly across both deterministic QE and probabilistic AI assurance — the “one estate” narrative is the core sell, and the architect has to embody it, not just recite it.

- Familiarity with the regulatory and standards backdrop shaping AI assurance conversations: EU AI Act, NIST AI RMF and ISO/IEC 42001.

📌 AI Architect (Pune)
🏢 Zensar Technologies
📍 Pune