Job posts for this role ask most for Python, with Scikit-learn, NumPy and Pandas for classic models, feature engineering and statistics to shape the data, and computer vision, speech processing and reinforcement learning for the specialised work. The full story of what each skill is used for is on the skills page.

Big tech is where machine learning work is most concentrated. Google, with Google DeepMind, is by far the largest employer of machine learning engineers in India. Most of its work is speech and language, recognising and generating speech and understanding text across Indian languages, and it is one of the most open places in the market for a junior (who they hire). Amazon hires a broad mix, classic models for its retail site, computer vision, and speech for Alexa, mostly with experienced engineers (who they hire). Microsoft and Apple have only a few openings.

Among product companies, machine learning is hired either to make machines see and sense, or to predict something inside a product. It is more open to juniors here than in most roles.

Industrial tech and mobility is the biggest hiring hotspot, and most of its work is computer vision and small models that run on devices. Across automotive and mobility, HARMAN builds automotive infotainment systems, with cameras that watch the driver and the road, Lytx builds fleet dashcams that spot risky driving, and Alstom and Wabtec build vision for trains and signalling. Among the chip makers, NXP Semiconductors, Microchip Technology and Analog Devices, which are analog, power and microcontroller chipmakers, hire engineers to fit small models onto their chips. They are among the most open employers in the market for a junior. NVIDIA, AMD, Qualcomm and Intel, processor and GPU makers, hire them to tune models for their chips. In energy and climate tech, Sustainability Economics.ai, GE Vernova and Enphase Energy build forecasting models, and FluxGen, which works on carbon accounting and sustainability software, takes freshers. GE Aerospace and Boeing hire a few for vision and maintenance on aircraft.

Prediction inside a product is the other hiring hotspot. The apps built for OTT and live streaming platforms, like Roku, Fubo, JioStar and JioHotstar, use recommendations, ad targeting and video understanding. Kaleris builds software for transportation management (TMS) and digital freight marketplaces, forecasting and routing freight, and Blue Yonder and Blubirch plan stock and returns, almost all with classic models (who they hire). Among e-commerce and D2C brands, eBay, Mercari, Myntra and Poshmark build models for search ranking, recommendations and pricing. ModMed, athenahealth and Tricog Health, which build software for digital health, hire engineers for models that read patient records and ECGs.

A few more pockets are worth knowing. Intuit's tax and personal finance products, part of stock trading apps and personal finance, use machine learning too. NetApp, Rubrik and Druva, which handle enterprise storage and backup, build models that spot ransomware. Zebra Technologies and Barco build vision into scanners and displays. The finance companies, the network and security companies and the developer-tool makers hire few machine learning engineers.

Among the GCCs, machine learning is used to predict risk, demand and prices for the parent.

The banks hire the most machine learning engineers of any GCC group. Inside the banks and financial institutions, NatWest Group, Standard Chartered, Goldman Sachs and Morgan Stanley build prediction models at the centres that run the bank's wider engineering, where machine learning is a real part of the work (who they hire). Inside the retailers and consumer-goods companies, Tesco, Walmart, Target and 7-Eleven build classic models for demand forecasting, pricing and promotions (who they hire). Inside the media, entertainment and publishing groups, Sony Research India hires a few. The healthcare, insurance, manufacturing and energy GCCs hire few or none. More on machine learning across the GCCs is in the GCCs' other roles.

The services world hires few machine learning engineers for the size of its hiring. The IT services giants, like Infosys, Capgemini, IBM and TCS, hire the most, for classic models in client projects. Quantiphi and Terawe, the AI and data specialists in the specialist firms built around one skill, hire a few, and TransPerfect, among the BPO and KPO firms, hires for speech and language data work. Accenture has openings too, though machine learning is a tiny part of its hiring (who they hire). The consulting firms' engineering arms hardly hire any. More on all of them is in the services world's smaller roles.

The hiring platforms, like Turing, Weekday AI, Uplers and hackajob, offer remote machine learning work for companies abroad, mostly with classic models, and they take juniors readily (who they hire). The staffing companies place few machine learning engineers.

Machine learning engineers are rarely hired by the payments, lending and banking software companies, the network and security companies, or the developer-tool and cloud companies. The GCCs of healthcare companies, insurers, manufacturers, logistics firms, energy majors and telecom operators hire few or none, and neither do the consulting firms' engineering arms.