← Careers·Forward Deployed Engineer
Engineering
Forward Deployed Engineer
About Comet Lab
Comet Lab puts AI to work inside a company's operations. The firm maps where AI fits within a business, then builds it: workflow automation, custom AI applications, and open source tools reshaped around how a team already works. Work spans client environments and Comet Lab's own products.
Role summary
The role is forward deployed. The engineer works inside client environments and on Comet Lab's own products, carrying an engagement from technical discovery through to systems running in production.
The founders define the high level architecture for each engagement, covering how information moves between departments, which workflows AI belongs in, and the role AI plays within each one.
This position owns everything downstream of that definition. It spans technical planning and feasibility assessment, the build itself, and the infrastructure the resulting systems run on.
Indicative allocation of time:
- 50 per cent technical planning and feasibility
- 25 per cent build
- 25 per cent infrastructure and deployment
The planning half carries the greater weight. A plan that has been properly tested makes the build straightforward.
This role is not suited to
- Candidates whose experience is confined to one layer of a larger team's system
- Candidates whose AI experience consists of prompting within no-code tools, with no code underneath
- Candidates who prefer to delegate the build and review the output
- Candidates who require a fully specified brief before starting
Terms
- Location: Mumbai, on-site, full time
- Working relationship: direct engagement with the founders, across client systems and Comet Lab's own products
How to apply
Applications should include a CV and a short note describing a system the candidate designed, built and deployed personally. The note should identify the point at which the initial plan proved wrong, and what changed as a result.
What you'll do
- Technical planning and feasibility: Convert a high level architecture into a concrete technical plan
- Technical planning and feasibility: Select the stack, including models, libraries, open source repositories, APIs and data stores
- Technical planning and feasibility: Stress test the plan before development begins, covering failure modes, scale limits and running cost
- Technical planning and feasibility: Challenge stated assumptions with supporting reasoning, and propose stronger alternatives where they exist
- Technical planning and feasibility: Produce the build plan that delivery then runs on
- Build: Develop the applications, integrations and agent workflows
- Build: Integrate third party systems over REST APIs, including the CRM, accounting and HR platforms already in use at the client
- Build: Apply AI coding tools as a standard part of the development process
- Infrastructure and deployment: Cloud infrastructure, servers, DNS, domains, load balancing, SSL and deployment pipelines
- Infrastructure and deployment: Take systems live, keep them running, and maintain visibility of failures
What we're looking for
- 3 to 6 years in software development, including production work on AI applications
- Demonstrated ownership of a system from concept through to production, rather than responsibility for a single layer of one
- Strong command of APIs, covering specifications, integration and diagnosis of responses
- Practical server experience across DNS, deployments, logs and certificates
- Well formed views on tooling, supported by reasoning
- The judgement to identify an unworkable approach early, and the confidence to raise it
- A habit of reading documentation, changelogs and library source
Nice to have
- Python, together with TypeScript or Node
- LLM applications in production, including retrieval, agents, evaluation and cost control
- Integration experience with business software such as Zoho, Salesforce, HubSpot or Google Workspace
Apply
Let's work together.
Fill out the form. Takes about 2 minutes. We'll be in touch if there's a strong fit.
Upload your CV (PDF or Word)
Optional: cover note & LinkedIn
We review with AI assistance and respond