GenAI Innovation Generalist

Keka Technologies Private Limited; Mumbai; On-site; INR 1,800,000 - 3,200,000; Full time

Job summary

Keka Technologies Private Limited seeks a GenAI-focused innovator who lives in the AI space. You will build fast demos, push prototypes, and decide what is hype versus where AI actually adds value, keeping the team ahead of industry shifts.

You will own the weekly AI radar, run prototypes to test ideas, craft the internal AI playbook, and lead workshops across teams to map AI opportunities to business goals.

Qualifications

• Fluency in AI concepts and business processes (sales, marketing, support, operations, finance).

Responsibilities

• Own AI radar of releases, tools, industry shifts and categorize into adopt/watch/skip.
• Build prototypes and demos to validate ideas before roadmap.
• Maintain internal AI playbook: tools, costs, best practices.
• Identify teams misusing AI and redirect.
• Lead tailored AI workshops for different teams.
• In first 90 days: ensure playbook adopted, ship 10+ reviews, build 2 prototypes, document what to stop doing.

Skills

RAG

Agents

Fine-tuning

LoRAs

Embeddings

Vector search

Context engineering

Open vs closed models

Job description

This is the most experimental role on the team. You live in the GenAI space. You test new tools the week they drop. You have real takes on what is overhyped and what actually matters. You can build a working demo before most people finish reading the launch blog post.

We want someone who keeps the whole team ahead, builds fast prototypes, and is not afraid to say this trendy thing is useless for us.

What You Will Own

• The weekly radar: AI releases, tool launches, and industry shifts, filtered into adopt, watch, or skip
• Prototypes and demos that prove or kill ideas before they become roadmap items
• The internal AI playbook: tools, policies, costs, best practices, kept current
• AI theater detection: teams using AI for show instead of impact get named and redirected
• Tailored AI workshops for different teams

Your First 90 Days

• The internal playbook exists and people use it without being told to
• 10+ adopt/watch/skip reviews shipped, each one an opinion with evidence, not a feature list
• 2 prototypes built fast enough to change a roadmap decision
• At least one documented “we should stop doing this” call, with what replaced it

What You Know

Not a tool list. Anyone can sign up for a vibecoding platform. This role is judged on two kinds of fluency and the ability to connect them.

The concepts, cold: RAG, agents, MCP, fine-tuning and LoRAs, evals, embeddings and vector search, context engineering, local inference versus cloud, open versus closed models. Not at paper-writing depth. At the depth where you know when each one is the right answer, when it is overkill, and what it roughly costs.

The business processes, equally cold: sales, marketing, customer support, operations, finance. How a lead gets qualified, how a support ticket flows, how a budget gets burned. Because the actual job is mapping the first list onto the second and being right about it. Someone who knows every model but cannot see where AI fits a lead-qualification flow is a hobbyist with opinions, and we already have Twitter for that.

Show Us

• Your top 10 AI follows on X and why
• A prototype or internal tool you built quickly to prove a point
• A take on something overhyped in AI right now and something underrated
• Experience killing an AI initiative that was not working and replacing it with something that did