Solutions Architect- LLM and Agentic AI

gnani.ai; Bengaluru; On-site; INR 600,000 - 1,200,000; Full time

Job summary

Gnani.ai builds voice AI for India, creating in-house speech models, Evon language models, and agentic AI platforms across 40+ Indian languages. This role is the technical face of Gnani's LLM and agentic AI work, shaping prototypes and POCs for large deals and representing Gnani to customers and partners.

As a hands-on leader, you will build and benchmark solutions, explain results, and engage with C-suite leaders to discuss trade-offs, latency, safety, and memory management, guiding proposals

Qualifications

• 8–12 years in tech with at least 3 years hands-on with LLMs and agent systems.
• Deep understanding of LLMs: architectures, training, fine-tuning, evals and benchmarks.
• Built and shipped agent systems in production with tool calling, memory, and multi-agent setups.
• Clear on trade-offs: large vs small models, open vs closed, cost vs latency vs quality.

Responsibilities

• Build prototypes and run POCs on real customer workloads.
• Design end-to-end solutions for large enterprise deals and help take them to production.
• Benchmark models and explain results clearly to non-technical stakeholders.
• Engage with CTOs, CIOs, and business heads to discuss tech, reasoning, tools, memory, safety, and latency.
• Lead the technical side of proposals and RFPs.
• Speak at conferences, panels, and webinars; write technical explainers.
• Provide market/customer insight to product and AI teams.
• Coach sales teams on Gnani's technology.

Skills

LLM architectures

Agent systems

Tool calling

RAG

Memory management

Stakeholder communication

Public speaking

Pre-sales support

Tools

Tool calling

RAG

Memory systems

Job description

Gnani.ai builds voice AI for India. We make our own speech models, language models (the Evon family), and agentic AI platforms that work across [40+] Indian languages. Banks, insurers, and telecom companies use our AI agents to handle millions of customer conversations. In August 2026, we launched Gnani Artha, India's sovereign AI stack, and open-sourced Evon 3.3.

About the role

You will be the technical face of Gnani's LLM and Agentic AI work. You understand how language models work from the inside, and you have built agents on top of them. You will prove our tech with prototypes and POCs, design solutions for large deals, and represent Gnani to customers, partners, and the industry.

This is a hands-on role. You build and benchmark things, then explain them.

What you will do

• Build prototypes and run POCs on real customer workloads.
• Design end-to-end solutions for large enterprise deals and help take them to production.
• Benchmark and evaluate our models against alternatives, and explain the results clearly.
• Sit with CTOs, CIOs, and business heads. Answer deep questions on how our agents reason, call tools, manage memory, stay safe, and keep latency low.
• Lead the technical side of proposals and RFPs.
• Speak at conferences, panels, and webinars. Write blogs and technical explainers.
• Bring customer and market insight back to the product and AI teams.
• Coach sales teams on our tech.

What we are looking for

• [8-12] years in tech, with at least 3 years hands-on with LLMs and agent systems.
• Deep understanding of LLMs: architectures, training and fine-tuning, evals, and benchmarks. You know what makes a model good or bad, and why.
• Built and shipped agent systems in production: tool calling, RAG, memory, multi-agent setups, guardrails.
• Clear on trade-offs: large vs small models, open vs closed, cost vs latency vs quality.
• Can explain hard ideas simply, to technical and non-technical people.
• Calm and clear when customers push back or compare us with others.
• Comfortable on stage.
• Customer-facing experience in a solutions, pre-sales, consulting, or founder role.

Good to have

• Experience with BFSI, insurance, or telecom customers.
• Voice AI or contact centre background.
• Talks, papers, blogs, or open-source work in AI.
• Working knowledge of data and AI rules such as DPDP.
• Work on India's own AI stack, built in-house from speech to language models to agents.
• Your hands on the models and platforms, not just the slides.
• See your work used at large scale by some of India's biggest companies.
• Direct access to product and AI leadership, and a real say in what we build next.