AI Solutions Engineer - Voice AI

About GoKwik
GoKwik is a growth operating system designed to power D2C and eCommerce brands—from checkout optimization and reducing return-to-origin (RTO), to payments, retention, and post-purchase engagement.
Today, GoKwik enables 15,000+ merchants globally, processes ~$2 billion in GMV, and continues to strengthen its AI-powered infrastructure. Backed by RTP Global, Z47, Peak XV, and Think Investments—and bolstered by a $13M growth round in June 2025 (total funding: $68M)—GoKwik is scaling rapidly across India, the UK, Europe, and the US.
Why This Role Matters
Voice is a new surface on top of GoKwik's stack: the agent that picks up when a merchant's customer calls. Every request that comes in is framed around one merchant's problem. This role's job is to solve that problem and turn it into a pattern, so the next merchant is faster to onboard than the last. Prompt quality is not a formatting detail here — on a voice call, the prompt is most of what the customer experiences, so this role directly shapes call outcomes and merchant trust in the product.
What You'll Own
Prompt Engineering & Agent Behaviour
Build and tune system prompts per merchant — structure the prompt, manage the variables inside it, control LLM behavior through instruction alone
Write for voice, not chat: sentence length, Hinglish switching, and how the TTS engine reads punctuation all change how a prompt performs out loud
Treat prompt hygiene as the main lever on call quality, not a downstream cleanup step
Merchant Solutioning & Pattern Design
Take a merchant's specific ask, stated or implicit, and design the conversation flow for it
Abstract the one-off into a reusable prompt template once the flow works, through pattern recognition across merchants
Resist starting with the integration just because it looks like the harder problem — the flow comes first, the generalization comes second
Data & Integration
Wire up product data and discount-variable flows through Shopify and third-party REST APIs, so the agent is working from clean, reliable inputs
Write light backend scripts for pre-processing: variable localization, price computation, and similar logic
Treat data reliability as a call-quality problem, not a separate engineering concern — a well-written prompt on bad data still fails the call
Continuous Improvement
Run test, listen, iterate as an ongoing loop, not a one-time deliverable
Listen to live calls to find where the agent breaks, then fix the prompt or the data behind it
Watch for the real risk: not that agents sound robotic, but that every merchant customization stays a one-off nobody can maintain, so the backlog grows faster than the team can close it
Who You Are
1–3 years of experience in a technical, product, or solutions role
Technical background helps but is not required — what matters is fluency with prompting, comfort building with AI tools, and enough understanding of how data moves through an LLM-plus-voice pipeline to connect it to business context
Strong prompting instincts: can structure a system prompt, manage variables inside it, and steer model behavior through instruction alone
Comfortable with ambiguity and iteration — this is a build-as-you-go role, not a spec-and-execute one
Why GoKwik
At GoKwik, you'll build one of the first voice AI products inside a fast-scaling eCommerce infrastructure company. You'll own agents end to end — prompt, data, and iteration — rather than handing pieces off to someone else. This role offers direct ownership of a new product surface, close contact with merchants, and the chance to set the pattern the rest of the team builds on.About GoKwik
GoKwik is a growth operating system designed to power D2C and eCommerce brands—from checkout optimization and reducing return-to-origin (RTO), to payments, retention, and post-purchase engagement.
Today, GoKwik enables 15,000+ merchants globally, processes ~$2 billion in GMV, and continues to strengthen its AI-powered infrastructure. Backed by RTP Global, Z47, Peak XV, and Think Investments—and bolstered by a $13M growth round in June 2025 (total funding: $68M)—GoKwik is scaling rapidly across India, the UK, Europe, and the US.
Why This Role Matters
Voice is a new surface on top of GoKwik's stack: the agent that picks up when a merchant's customer calls. Every request that comes in is framed around one merchant's problem. This role's job is to solve that problem and turn it into a pattern, so the next merchant is faster to onboard than the last. Prompt quality is not a formatting detail here — on a voice call, the prompt is most of what the customer experiences, so this role directly shapes call outcomes and merchant trust in the product.
What You'll Own
Prompt Engineering & Agent Behaviour
Build and tune system prompts per merchant — structure the prompt, manage the variables inside it, control LLM behavior through instruction alone
Write for voice, not chat: sentence length, Hinglish switching, and how the TTS engine reads punctuation all change how a prompt performs out loud
Treat prompt hygiene as the main lever on call quality, not a downstream cleanup step
Merchant Solutioning & Pattern Design
Take a merchant's specific ask, stated or implicit, and design the conversation flow for it
Abstract the one-off into a reusable prompt template once the flow works, through pattern recognition across merchants
Resist starting with the integration just because it looks like the harder problem — the flow comes first, the generalization comes second
Data & Integration
Wire up product data and discount-variable flows through Shopify and third-party REST APIs, so the agent is working from clean, reliable inputs
Write light backend scripts for pre-processing: variable localization, price computation, and similar logic
Treat data reliability as a call-quality problem, not a separate engineering concern — a well-written prompt on bad data still fails the call
Continuous Improvement
Run test, listen, iterate as an ongoing loop, not a one-time deliverable
Listen to live calls to find where the agent breaks, then fix the prompt or the data behind it
Watch for the real risk: not that agents sound robotic, but that every merchant customization stays a one-off nobody can maintain, so the backlog grows faster than the team can close it
Who You Are
1–3 years of experience in a technical, product, or solutions role
Technical background helps but is not required — what matters is fluency with prompting, comfort building with AI tools, and enough understanding of how data moves through an LLM-plus-voice pipeline to connect it to business context
Strong prompting instincts: can structure a system prompt, manage variables inside it, and steer model behavior through instruction alone
Comfortable with ambiguity and iteration — this is a build-as-you-go role, not a spec-and-execute one
Why GoKwik
At GoKwik, you'll build one of the first voice AI products inside a fast-scaling eCommerce infrastructure company. You'll own agents end to end — prompt, data, and iteration — rather than handing pieces off to someone else. This role offers direct ownership of a new product surface, close contact with merchants, and the chance to set the pattern the rest of the team builds on.

Work arrangement
No