How to hire a forward deployed engineer (AI) in Mumbai
Start here
What does a forward deployed engineer do with AI products?
A forward deployed engineer (sometimes called a forward deployment engineer or FDE) works directly with customers to deploy, connect and adapt an AI-powered product to the customer own systems and workflows. The product may be built on large language models, and the customer wants it to work on their data, in their way.
The job is part engineering, part customer work. Decide which of these parts you need.
Scope the problem
Sit with the customer, find the real need and agree what to build first.
Integrate
Connect the product to the customer systems, files and data sources.
Set up the AI
Prompts, retrieval over the customer documents and small custom tools or agents.
Check and report back
Test output quality on the customer real data and pass learnings to the product team.
Skills
Skills and tools to ask for
This person needs coding, AI know-how and clear talk with customers together. A gap in any one shows quickly at a customer site.
Python or TypeScript
Writing clean scripts, integrations and small tools.
APIs and integrations
Calling APIs and connecting systems that were not built to talk to each other.
LLM APIs
Using language model APIs, with an eye on cost and response time.
Prompt design
Clear instructions and examples that give steady results.
Retrieval over documents
Finding the right pieces of company documents and giving them to the model.
Evaluating AI output
Test examples from the customer data to see if answers are right.
SQL and data handling
Reading, cleaning and joining customer data.
Cloud and deployment basics
Running the work on a server or cloud, with logs to find problems.
Talking to customers
Explaining choices to non-technical people in short, plain words.
Problem scoping
Turning a vague request into a small first step that can be tested.
Also important:
- Careful handling of customer data: asking what may be shared with an outside model, and removing what should not be.
- Honesty about what the AI can and cannot do, rather than over-promising.
- Short, clear notes on what was set up and why.
- A steady manner when a customer is unhappy.
The job description
What to put in the job description
Be honest about how much time is at customer sites and how much is coding. Keep it short and include these points. More on this in how to write a job description.
- The title and the office area, plus how often the person is at customer sites.
- What your product does and what a typical customer set-up looks like.
- The languages and tools you use, for example Python, SQL and Docker.
- Minimum experience, and whether you will consider someone newer with strong projects.
- Rules for customer data, and travel needs with who pays.
- Your pay range, to be given honestly on the first call.
- Title
- Forward Deployed Engineer (AI)
- Location
- Our office in [your area], Mumbai, with regular visits to customer sites.
- Experience
- 2 to 5 years in software, with at least one project using language model APIs. Customer-facing work is a plus.
- Skills
- Python or TypeScript, REST APIs and integrations, language model APIs, prompt design, retrieval over documents, SQL, Git, cloud basics. Clear communication with non-technical people.
- The work
- Understand each customer problem, set up our AI product on their systems and data, build small integrations or tools, test output quality on their real data, handle their data with care, train their staff and report what we learn to the product team.
- Hours
- [Days and hours]. Site visits may need early or late days.
- Salary
- [Your range], to be discussed on the call.
- To apply
- Tell us about one customer problem you solved. We will call you if your profile matches.
Check
How to check real work and customer skills
A resume can show the coding side. The customer side you have to see for yourself.
- Ask them to walk through a past customer deployment: the problem, what they built, what went wrong and what they changed.
- Give a short practical task of an hour or two, such as connecting a language model API to a small set of sample documents, and ask what trade-offs they made.
- Ask how they checked that the answers were good on real data.
- Ask them to explain one technical choice to you as if you are not technical.
- Play a short role: you are an unhappy customer whose data will not load. Watch how they ask questions.
- Ask what they did when a customer asked for something the product could not do.
Use our first-call questions for the first call. If what you really need is someone to build the AI product itself, read how to hire a GenAI developer. The GenAI developer builds the product. The forward deployed engineer adapts and deploys it at customers.
Where
Area, travel and hybrid work in Mumbai
Because this person moves between your office and customer sites, where they live matters more than for most roles. Mumbai travel is slow at peak hours.
Think about where your customers are. Search near your office first, then along the same train or metro line. See hiring by Mumbai locality.
Search in plain words, for example “Python engineer with language model API and customer-facing work in Andheri”. See how to use the resume database. A related read is how to hire a software engineer.
Avoid these
Common mistakes when hiring a forward deployed engineer
Mistake
Hiring only for AI skills
The job fails if the person cannot talk to the customer.
Better
Test communication as seriously as code.
Mistake
A vague title with no detail
Candidates will expect different jobs.
Better
Say how much is customer time and how much is building.
Mistake
Judging by an AI demo
A demo can hide wrong answers.
Better
Ask how they checked quality on real data.
Mistake
Ignoring customer data rules
Better
Ask what they would not send to an outside model. Check with a professional about the rules for your business.
Mistake
No one to back them up
Better
Name the person in your team they can ask when stuck.
FAQ
Common questions
What is a forward deployed engineer?
A forward deployed engineer (sometimes called a forward deployment engineer or FDE) works directly with customers to deploy, connect and adapt an AI-powered product to the customer systems and daily work. They write real code, but they also spend time with the customer, understand the problem and report what they learn back to the product team.
How is this different from a GenAI developer, a solutions engineer or a support engineer?
A GenAI developer builds the product itself. A forward deployed engineer takes that product to a customer and makes it work there. A support engineer mostly fixes problems after launch, and a solutions engineer often focuses on demos and pre-sales. In small teams one person may do parts of several of these, so write down which parts you need.
Do they need to be senior?
Not always, but they must be able to work without close supervision, because they are often alone in front of a customer. Look at whether they have handled a messy customer problem on their own, more than at the number of years.
Will they travel or work at customer sites?
Usually some of the time, to customer offices in Mumbai or elsewhere. Some work can be done remotely. Say clearly in the job description how many days are at customer sites and who pays for travel.
Related reading
- Resume database for recruiters
- Pricing
- Hiring by Mumbai locality
- Search from a job description
- Hire an accountant
- Hire telecallers
- Hire sales executives
- Hire a receptionist
- Hire warehouse staff
- Hire a GenAI developer
- Hire a software engineer
- Hire a DevOps engineer
- Hire a data analyst
- Hire an Android developer
- Hire an iOS developer
- Hire a full stack developer
- Hire a QA engineer
- Shortlist without HR
- First call questions
- Check experience
- Set a fair salary
- Fresher or experienced
- Fill an urgent vacancy
- Replace someone who quit
- Seasonal hiring
- First-week onboarding
- Offer letter basics