Forward Deployed Engineer (FDE) — AI & Platform Solutions

Sukrut Associates Company; Jaipur; On-site; INR 250,000 - 500,000; Full time

Job summary

Sukumrut Associate in Jaipur, India, seeks a Forward Deployed Engineer to sit inside the client's problem, designing and delivering AI tooling for healthcare, legal and fintech sectors. You own systems end-to-end—from architecture to production support—and work directly with clients to ensure the solution actually works.

You will use Claude Code and Cursor in the loop, with self-hosted models when required, and integrate with live systems. US-hours calls from IST are expected.

Qualifications

• 10+ years of engineering experience.
• Able to own architecture from design to production support.
• Willing to work US business hours from India.
• Experience with AI tooling and multi-service platforms.

Responsibilities

• Design, build and maintain AI tooling and multi-service platforms for clients.
• Own systems end-to-end—from architecture through production support.
• Work directly with clients, with no intermediary translation of problems.
• Collaborate in a US-facing schedule and ensure client satisfaction.

Skills

End-to-end ownership

Client-facing collaboration

AI tooling experience

Job description

Most engineering roles end at the pull request. This one starts there.

As a Forward Deployed Engineer at Sukrut Associate, you sit inside the client's problem — a legal team drowning in medical-record review, a healthcare finance operation reconciling payment determinations by hand — and you build the thing that makes it disappear. You design it, ship it, and then stay with the people using it until it actually works.

We build custom AI tooling and multi-service platforms for clients in healthcare, legal, and fintech: document-intelligence pipelines, fine-tuned LLM workflows, Salesforce-native applications, and the AWS infrastructure underneath all of it. You will own systems end to end — architecture through production support — with direct access to the client and nobody translating the problem for you.

You will work the way we work: Claude Code and Cursor in the loop daily, open-weight models (Llama, Qwen, Mistral, DeepSeek, Hermes) fine-tuned and self-hosted when the client's data demands it, and MCP servers wiring our agents into real systems. We care that you use the sharpest tool available, not that you memorised a stack.

One practical note: our clients are US-based. You will need reliable availability for live calls in US business hours (ET/PT), which means regular evening and late-evening IST meetings.

This is a role for someone 10+ years in who is tired of implementing specs handed down from three layers up. If you like being the person in the room who both understands the client's workflow and can write the code that fixes it, we should talk.