About the Role

Forward Deployed Engineer Role Type: Full-time | Individual Contributor (Hands-on) Experience: 2-3+ years overall, with hands-on delivery using LLMs and AI development tools The Forward Deployed Engineer closes the distance between an AI prototype and a system a team depends on in production. You are an engineer who works from inside the reality of the team you're embedded with — not from behind a ticket queue. You sit with the people who have the problem, learn how they actually operate, and build the thing that fixes it: AI agents, integrations, and the services behind them.

Most of what you build connects real systems to each other, under constraints you do not control — interfaces you have to work inside, data that is rarely as clean as the documentation suggests, and a request that describes the symptom rather than the problem. Your job is to move quickly inside those constraints and put something working in front of real users. You are embedded with Bonterra's own business and product teams rather than with external clients, so you work with the same stakeholders over months and build real depth in their domain.

Then you ship it, watch real people use it, and make it better. You're measured by what runs in production and what it changes, not by tickets closed.

Responsibilities

- Embed with a team or function, learn how they work, and build the agents and integrations they need
- Take part in discovery — sit with users, watch the workflow as it actually runs, look at the real data, and help turn what you learn into a scoped piece of work with a clear definition of done
- Build integrations end to end against third-party platforms — authentication, pagination, rate limits, error handling, and the data model underneath
- Write production Python for backend services and AI workloads, including LLM integration, tool calling, and retrieval
- Move and reshape data between systems using SQL and pipelines, and understand the data well enough to know when it is wrong
- Work with AI coding tools as a daily part of the job,



and review what they produce before it ships
- Ship fast, then iterate in the field — prototype in days, get it in front of real users, and improve it based on what actually happens in production
- Stay with what you ship after go-live — watch the logs, the traces, and the user feedback, and fix what breaks
- Handle the things that normally derail timelines — incomplete data, workflow exceptions, missing documentation — by unblocking rather than escalating
- Flag recurring problems so we build the solution once and reuse it rather than solving it again for the next team
- Take a vague request from a non-technical stakeholder and turn it into a scoped, shippable piece of work
- Grow into ownership of larger surfaces as you build depth across our platforms

Qualifications

- 2-3+ years of professional software engineering experience, with real delivery rather than coursework or training projects

Required Skills

- Solid hands-on Python, written daily and comfortably debugged in somebody else's codebase
- Ability to read unfamiliar code and explain what it does — this is assessed directly in the interview and matters as much as writing it
- Solid fundamentals in data modeling, authentication and authorization flows, REST APIs, pagination, rate limiting, retries, and asynchronous or concurrent execution
- Working SQL, and comfort reasoning about how data is structured across systems
- At least one integration or backend service shipped to production that you can walk through in depth — why you designed it that way, what went wrong, and what you would do differently
- Demonstrated delivery using LLMs or AI coding tools, with informed opinions on where they help and where they do not




- Working familiarity with LLM application patterns — prompting, tool use, and retrieval — at the level of having built something with them
- Exposure to at least one major cloud platform (AWS, GCP, or Azure) and to deploying something you built
- Comfort with version control, testing, and debugging in a shared codebase
- Able to explain a technical constraint to someone non-technical without losing them

Preferred Skills

Nice-to-Have We hire for fundamentals and speed of learning rather than tool checklists. None of the following is a filter — if you have them, good; if not, you will pick them up here.

- AI frameworks such as LangChain, LangGraph, or LlamaIndex
- RAG in depth — chunking strategies, embedding pipelines, vector databases, and retrieval evaluation
- Hands-on AWS. We run on AWS; experience on GCP or Azure transfers and is equally welcome.
- Docker, CI/CD, and infrastructure-as-code
- Non-relational databases (DynamoDB, MongoDB, Redis, vector stores, or similar)
- Evaluation frameworks, observability, guardrails, and cost controls for AI systems
- Node.js or TypeScript for service development and integrations
- Frontend familiarity (React, Next.js) for putting a usable interface on AI tools

What We're Looking For (Attributes)

- Builder mindset: You ship. You'd rather have something imperfect in front of a user this week than something perfect in a month.
- High agency: Dropped into an unfamiliar team with a fuzzy problem, you go figure it out — you don't wait for a spec.
- Asks what the problem is for: Before designing anything, you find out what the output is actually used for and who uses it. This is the single most important habit in the role.
- Comfortable without a spec: You can make real progress when the requirements are incomplete and still changing.
- Pragmatic about scope: You can tell the difference between what is needed and what is impressive, and you build the first one.
- Customer-embedded, not code-only: You can sit with a non-technical

📌 Forward Deployed Engineer (Hyderabad)
🏢 Bonterra
📍 Hyderabad