Information Technology & Software
18 Nov
Solutions Architect — Presales (Intelligence Layer & AI Factory, Sovereign AI)
Solutions Architect — Presales (Intelligence Layer & AI Factory, Sovereign AI) Location: Mumbai, India (travel to client sites — government and enterprise — across India) Function: Presales / Sales Engineering, AI & HPC Infrastructure Reports to: Presales / Sales Leadership About the role The organisation designs and delivers large-scale sovereign and enterprise AI solutions across India — from GPU infrastructure and AI software platforms through to the AI Factory (MaaS/GPUaaS) and the use cases running on top of it. This role focuses on the intelligence layer and above: the AI Factory (Model-as-a-Service, GPU-as-a-Service), the platform services that sit above raw compute, and the use-case layer where those services become a solution a government or enterprise customer can adopt. It covers two related areas of work: architecting and commercially structuring AI Factory (MaaS/GPUaaS) offerings, and owning the solution design for full turnkey sovereign AI programs for large government customers, spanning every layer of the stack. What's expected • Solutioning at the intelligence and use-case layers — model serving strategy, RAG/agentic pipeline design, governance/guardrails, and translating customer use cases (departmental, citizen-facing, defence, industryspecific) into a concrete solution built on the AI Factory. • AI Factory (MaaS/GPUaaS) presales — sizing and architecting Model-as-a-Service and GPU-as-a-Service offerings, including multi-tenant/dedicated capacity tradeoffs, and building the commercial structure (pricing, margin, tenancy/SLA/security models) that makes them sellable, whether the customer is consuming capacity or reselling it. • Full turnkey sovereign AI proposals for large government customers — pulling together the complete stack into one coherent architecture and narrative, and owning the intelligence-layer-up design and integration. • Financial and commercial modeling for AI Factory and turnkey deals — BOMs, pricing, margin models, and, where relevant (PPP/concession structures, investment-grade government programs), IRR/NPV/investmentjustification models suitable for a government finance department or investor audience. • Technical presales for sovereign and government engagements — bilateral state-government propositions, large public-sector procurement responses, and multi-stakeholder programs — often requiring policyliterate framing (PPP structures, procurement/regulatory instruments) alongside the technical design. • Contributing to internal tooling — the practice maintains a growing internal suite of sizing, pricing, and lifecycle-modeling tools, including AI Factory/AIaaS configurators, financial engines, and sovereignty/procurement scoring frameworks. This role is expected to help extend them, especially at the intelligence/use-case layer. • Proposal and pitch ownership — decks, solution architecture documents, RFP responses, and executivelevel pitch material, tailored to audiences from a state government's leadership down to a customer's technical evaluators. • Delivery handoff — handing off a clean, layered solution to systems integration/delivery, with clear ownership boundaries against adjacent teams. • Building reusable IP, not one-off answers — every engagement should leave behind a template, model, or framework (an AI Factory sizing pattern, a use-case solution blueprint, a turnkey proposal structure) that makes the next one faster. What's not expected • Infrastructure-layer sizing — GPU cluster architecture, networking, datacenter/cooling design, and similar sit with a separate infra/orchestration presales function; this role integrates with that layer rather than owning it. • Orchestration-layer design — workload scheduling, cluster orchestration, and related platform-belowintelligence concerns are likewise owned elsewhere. • Third-party public cloud architecture — this is a sovereign/on-prem and AI Factory-focused role, not a public-cloud (hyperscaler) solutions role. What we're looking for • 10–15 years of overall technology/solutions experience, including 3–4 years specifically in AI — this is a senior, customer-facing solutioning role, not an entry point into AI. • Strong grounding in the intelligence and use-case layers of an AI stack — model serving and inference patterns, RAG/agentic architectures, AI platform services (data fusion, model registry, APIs/gateways, guardrails/governance), and enough fluency in the underlying infra/orchestration layers to integrate cleanly with the team that owns them. • Direct experience with AI Factory-style offerings — Model-as-a-Service and/or GPU-as-a-Service, including the commercial and tenancy models (multi-tenant vs. dedicated capacity, SLAs, security tiers) that make them viable products, not just technical demos. • Fluency in token economy, sizing, and inference optimisation — token-based consumption/pricing models, throughput and latency sizing at the model-serving layer, and optimisation levers (quantisation, batching, speculative decoding, caching, and similar) well enough to architect and defend a solution on these terms, not just quote a vendor number. • Ability to stand up a sovereign cloud offering end-to-end — comfortable with everything that goes into building a sovereign AI cloud for a government customer, from CoE/government-use and commercial-cloud track structuring through to the day-to-day mechanics of running it as a service — the full scope a large bilateral sovereign AI program requires, not just the technical solution slice. • Comfort with financial and commercial modeling — margin-based pricing, TCO, and ideally IRR/NPV/investment-justification modeling; able to hold a real conversation with a finance department, not just an engineering team. • Government/public-sector experience is a strong plus — familiarity with how public procurement, PPP/concession structures, and multi-stakeholder government proposals actually work, given the turnkey sovereign AI focus of much of this book of business. • Ability to own a full-stack narrative without owning every layer — comfortable being accountable for a turnkey government proposal while relying on a separate infra/orchestration presales function for that portion of the design. • Builder mindset toward tools, not just decks — someone who'd rather build (or push for building) a proper sizing/solution framework than maintain a spreadsheet full of manual guesses. • Excellent written and presentation communication — able to flex between a technical solution review and a strategic pitch to senior government or enterprise stakeholders. • Willingness to travel across India for client engagements, government stakeholder meetings, and site work. Nice to have • Prior exposure to sovereign AI, national compute/AI infrastructure programs, or public-sector digital infrastructure. • Experience selling or architecting AI Factory-style commercial models (MaaS/GPUaaS/BMaaS) specifically, rather than general cloud services. • Familiarity with defence, geospatial, or other classified/regulated use-case domains. • Experience contributing to or shaping internal tools/platforms (not necessarily as a developer — as a domain expert who drives requirements)