Kaya AI
Website:
kayatech.com
Company:
https://www.linkedin.com/company/usekayaai
Seniority: Mid-Senior level
Industries: Software Development
Job details:
Head of Kaya Labs, India
Location: Mumbai or Bangalore preferred, remote within India considered, with regular travel to partner campuses
Employment type: Full-time
Experience: 5+ years, with at least 3 years hands-on in ML or LLM systems
About Kaya AI
Kaya AI is building the system of action for mission-critical construction and infrastructure projects. We help owners, developers, and contractors proactively manage procurement, logistics, and execution across some of the most complex projects in the world, including data centers, life sciences, and advanced manufacturing.
Our platform combines AI, workflow automation, supply chain intelligence, and real-world operational data to help teams deliver projects faster and with greater predictability. Our agentic coworker, Amber, learns customer context in minutes and helps teams complete RFPs faster, win more work, and eliminate hours of manual procurement and logistics work every week.
We are backed by leading investors and trusted by industry leaders including Crusoe, Align Data Centers, Accenture, Suffolk Construction, and Turner. Kaya has been recognized by Infrastructure Masons (IM 100), BuiltWorlds as a Top AI-Driven AEC Solution for 2026, Cemex Ventures Top 50 Contech Startups, and Suffolk Technologies as Startup of the Year.
About Kaya Labs
Kaya Labs is Kaya's applied research function. Construction procurement and supply chain workflows are high-stakes, data-heavy, and built on fragmented, largely unstructured inputs: specs, quotes, submittals, drawings, and schedules. Foundation models are getting cheaper and more capable, which shifts the advantage to whoever has the domain-specific models, evaluation sets, and learning loops. Kaya Labs exists to build those.
Labs builds prototypes, datasets, and reference implementations that Kaya's product engineering team productionizes. It works with universities, professors, and student builders to move faster and to grow a long-term talent pipeline. It is a small, focused track, not a parallel feature team and not a broad research organization.
The Role
We are hiring the founding lead for Kaya Labs in India. This is a hands-on applied AI research role with a program-building mandate attached. You will personally build the first generation of Kaya's domain-specific models, and you will build the university partnerships and intern pipeline that let Labs scale beyond one person.
In the first year, Labs is focused on three things: a procurement entity and intent model that extracts and normalizes materials, quantities, suppliers, lead times, and constraints from real project documents; a supplier and lead-time intelligence model that flags delivery, price, and substitution risk; and a domain retrieval system that improves grounded search and Q&A over specs, drawings, quotes, and submittals. You will own the technical direction of this work, the datasets and evaluation harnesses behind it, and the handoff of working outputs to product engineering in the US.
You will report to [the CEO / Director of AI Engineering] and work in close partnership with Kaya's engineering and product leadership. You will be the interface between the work happening in India and the team shipping it to customers, so regular overlap with US Pacific hours is expected.
What You'll Work On
Build domain-specific models for procurement and supply chain: information extraction, entity normalization, risk scoring, and retrieval over construction documents
Create labeled datasets and evaluation harnesses that become reusable Kaya assets and run as weekly regressions
Deliver prototypes and reference implementations with clear accuracy numbers and failure modes, built to product engineering's constraints so they can be productionized
Establish and run applied research collaborations with professors and labs at IITs, IISc, BITS, and similar institutions, under governed and anonymized data agreements
Recruit, scope, and manage an ongoing intern and student-builder pipeline, and mentor those teams to working prototypes
Plan hackathons, campus initiatives, and showcase events that feed the pipeline and build Kaya's reputation in the Indian AI community
Set quarterly Labs priorities with Kaya leadership and own the handoff of integration-ready outputs to product engineering
What Success Looks Like
In the first six months: working prototypes for at least two of the three focus areas, a labeled procurement dataset and eval harness in regular use, one output that product engineering has picked up for production, and at least one university collaboration underway.
In the first year: at least one Labs output embedded in production workflows with a measurable improvement in procurement accuracy, speed, or risk detection, and a repeatable model for working with one or two university partners.
Ideal Candidate
5+ years of experience, including at least 3 years hands-on building ML or LLM-based systems that shipped
Strong applied depth in at least two of: information extraction and NLP, retrieval and RAG, evaluation design, knowledge graphs or structured data modeling, predictive modeling on operational data
Has built training and evaluation datasets from messy real-world documents and knows what good labeling and measurement look like
Existing relationships with faculty and student communities at leading Indian technical institutions, or a credible plan to build them
Experience mentoring or managing early-career engineers and getting real work out of student teams
Strong ownership and comfort operating with a wide mandate in an early-stage company
Clear written and verbal communication with engineers, professors, and executives across timezones
Preferred Technical Experience
Python and the modern ML and LLM toolchain, including fine-tuning, embeddings, and vector search
Document AI and multimodal extraction over PDFs, drawings, and tables
Evals, observability, and quality systems for AI products
Postgres and cloud infrastructure on AWS, GCP, or Azure
Enough full-stack fluency to build a demoable prototype end to end
Bonus Points
Background in construction, supply chain, logistics, or industrial software
Prior experience running university partnerships, incubators, hackathons, or talent programs
Published research or public technical work in NLP, retrieval, or applied ML
Early or founding role at a venture-backed startup
Why Join Kaya Labs
Found and shape a research function from the ground up, including its program, partnerships, and talent network
Work on frontier AI applied to a massive, underserved real-world domain, deployed on hyperscale data centers and critical infrastructure projects
Direct partnership with Kaya's founder, leadership, and US engineering team
Real feedback loops from live projects and outcomes, which is what turns generic AI into durable domain advantage
Competitive compensation and meaningful equity in a category-defining company
To apply: Send your resume, a link to one thing you have built or one program you have run end to end, and a short note on why you want to build Kaya Labs in India.
Click on Apply to know more.