AI Jobs in India: 109 Live Roles and 11 Remote (September 2026)

India is the third largest AI job market on this board. On 8 September 2026 the India page listed 110 live AI roles, with the public stats API reporting 109 in the same read window, behind only the United States (1,803) and the United Kingdom (190) and ahead of Canada (72).

That headline is the least interesting thing in the data. What matters is the composition, and on three separate measures the Indian AI market looks structurally different from the American one: who is hiring, what they are hiring for, and what they are willing to say about pay. This edition is frozen at publication; the live list is always at AI jobs in India.

The employers are multinationals, not labs

The 50 most recent India listings on 8 September 2026 came from 31 different companies, and the ranking is dominated by global firms hiring into Indian capability centres rather than by AI companies: Adobe (5 roles), PwC (4), Bristol Myers Squibb (3), DevRev (3), then Novartis, Amgen, Wells Fargo, Micron, CrowdStrike, Databricks, Sarvam AI and Anthropic with 2 each. Single roles came from Takeda, NVIDIA, AstraZeneca, Roche, Visa, Autodesk, Equifax, Workday, S&P Global, State Street, Chevron, Thomson Reuters, Hewlett Packard Enterprise, Composio, Together AI, Nebius, Freehand, Anysphere (Cursor) and OpenAI.

The sector concentration is the surprise. Ten of the 50 listings were from pharma and life sciences: Bristol Myers Squibb 3, Novartis 2, Amgen 2, and one each from Takeda, AstraZeneca and Roche. A further six came from financial services and market data: Wells Fargo 2, plus Visa, Equifax, State Street and S&P Global. Anyone picturing the Indian AI market as startups in Bengaluru has the weighting backwards. It is regulated industry building internal AI capability, at scale, in India.

AI-native employers are present but thin on the ground. Anthropic had two Applied AI Architect roles in Bangalore, both hybrid. OpenAI had an Applied AI Engineer in Delhi, also hybrid. Sarvam AI, the domestic model builder, had an Agent Engineer and a Performance Engineer, Inference, both in Bengaluru. Those five roles are the frontier-lab footprint in India as of that read.

Bengaluru and Hyderabad, and then a long way down

Location as listed across the same 50 roles:

• Bengaluru or Bangalore - 18
• Hyderabad - 12
• India, no city given - 7
• Pune - 4
• Noida - 3
• Chennai - 2
• Gurugram - 2
• Delhi - 1
• Kolkata - 1

Two cities held 30 of 50. Hyderabad's second place is not a rounding artefact either: it is where the pharma capability centres sit, with Novartis, Amgen, Bristol Myers Squibb, Roche and Micron all listing there. Mumbai, often named alongside the others as an Indian tech centre, appeared in none of the 50.

It is applied work, and four functions are simply absent

Classified by function, the 50 roles split:

• Applied AI - 31
• ML Engineering - 9
• AI Infrastructure - 6
• Data & MLOps - 3
• AI Product - 1
• Research Science - 0
• AI Safety & Alignment - 0
• Evals & Red Teaming - 0
• AI Governance & Policy - 0

Applied AI alone is 31 of 50, roughly three roles in five, and the titles say why: Senior AI Engineer, AI Platform Engineer, Forward Deployed Engineer, Conversational AI Developer, Generative AI Solutions Architect. This is deployment work, integrating models into existing products and processes for an employer whose core business is something else.

The four zeroes are the finding a candidate should act on. Across those 50 listings there was no research science role, no safety or alignment role, no evals seat and no AI governance role. If that is the career you want, the current Indian market is not where it is being hired, and the realistic route is a remote or relocated role. The volume in India is at AI engineering and machine learning engineer.

Remote in India is more or less a single job title

Of the 378 fully remote AI roles on the board, 11 named India as the base country, behind the United States (211), Canada (22) and the United Kingdom (20). Eleven is a small enough number to read individually, and it is startlingly uniform: 8 of the 11 carried Forward Deployed in the title, and Databricks alone held 5 of the eleven. The rest were DevRev, CrowdStrike, Together AI, LiveKit, Level AI and OpenAI. Ten of the eleven were classified Applied AI and one AI Infrastructure.

The pattern is coherent: a fully remote India hire, at a company headquartered elsewhere, is usually someone deployed to customers rather than someone building the core system. The CrowdStrike listing states the trade-off out loud in its own title, scoping the role to 2.00 pm to 11.00 pm IST. Most postings leave that unsaid, which makes time-zone overlap the first thing to ask on a screening call. The live list is at remote AI jobs in India, and the board-wide picture is in the September remote edition.

Nobody publishes pay

None of the 50 India listings displayed a salary band. Across the 11 fully remote India roles, exactly one did: LiveKit's Forward Deployed Engineer, at 80L to 1.4Cr rupees, quoted as displayed. One published band out of 61 listings read. India has no pay-transparency statute of the kind that forces disclosure in several US states, so the absence is expected, but it is worth stating plainly: any Indian AI salary figure you read online is survey data or self-reporting, not employer disclosure, and it should be treated accordingly.

What to do with this

• Apply to capability centres, not just to AI companies. Adobe, PwC, the pharma group and the banks are the volume. AI-native labs had five India roles between them in this read.
• Lead with deployment evidence. 31 of 50 roles are Applied AI. Shipping a model into someone else's system beats a research portfolio for this market.
• Look at Hyderabad, not only Bengaluru. 12 of 50, and a different employer mix, with less competition than the Bengaluru pool.
• Treat forward deployed as the remote door. Eight of the eleven remote India roles were that title. It is the shape of remote AI work available from India today.
• Ask for the band, because it will not be posted. One in 61 listings published one. You have no external anchor unless you create it.

Method and caveats

Country counts were read 7 to 8 September 2026 from /api/public/stats and the country pages. Every company, city, function and pay figure above was read on 8 September 2026 from the 50 listings then showing on /ai-jobs-in-india and the 11 on /remote-ai-jobs-in-india. The 50 are the most recent listings, not the full 110, so the company and city shares describe recent hiring rather than the standing inventory. Cities are grouped where employers spell the same place differently (Bengaluru and Bangalore counted together); the seven roles giving only India are counted separately rather than assigned to a city.

Two limits apply to the totals. This is a large sample of the companies that publish an applicant tracking system feed, not a census, so treat every count as a floor. And enterprise Workday feeds are capped per tenant, which under-counts large multinationals. India has many of those, and they are exactly the employers this market runs on, so the true Indian figure is materially higher than 110. Roles are read directly from each employer's own applicant tracking system, and pay is quoted exactly as published. Every figure here is free to quote with attribution to AI Hiring Board.