AI Forward Deployed Engineer

Department: AI

Location: Remote - India

Description
About us:
Hiver is a modern, AI-driven customer service platform used by companies across healthcare, finance, logistics, education, and technology. We help teams deliver fast, human support across email, chat, phone, WhatsApp, and more — without the complexity of legacy helpdesks.
Were a challenger brand in a category dominated by over-engineered tools. We build software that is simple, powerful, and genuinely helpful, and we operate internally with that same philosophy. If you want meaningful ownership, thoughtful teammates, and work that ships, Hiver is a great place to do it.
Opportunity:
The AI Forward Deployed Engineer is part of the AI engineering team and drives adoption of Hiver AI. You embed (virtually) with strategic customer accounts, understand how their support and operations teams actually work, and then build — production-grade configurations, automations, knowledge pipelines, and integrations that make Hivers AI deliver measurable outcomes in their environment. You stay until it works, and you carry what you learn back into the product.
This is an AI engineering role, not a sales role. No quota, no commission, no demo circuit. Your success metric is whether the customers AI deployment is live, trusted, and adopted.
Key Responsibilities
What are we looking for

• 4–6 years as a software engineer, with at least 1–2 years working hands-on with LLM-powered systems in production: prompt and context engineering, RAG pipelines, agentic workflows, eval harnesses.
• Strong Python; comfortable with TypeScript/JavaScript for full-stack work (dashboards, integrations, internal tools).
• Real API/integration experience — youve connected messy third-party systems before and know that the documented behaviour and the actual behaviour are different things.
• Excellent written and verbal communication. You can run a discovery call with a non-technical support manager, explain a misclassification to a skeptical ops lead without jargon, and write a deployment doc someone else can follow.
• High ownership and comfort with ambiguity. There is no playbook for most of what youll do — youll write the playbook.
• Maintain work-hour overlap with Engineering, US-based teams, and customers to foster effective cross-functional collaboration.

What you will do

• Embed with strategic accounts. Join shared Slack channels, sit in on the customers team rituals, shadow real ticket queues, and map how work actually flows through their shared inboxes — all remotely.
• Build the last mile. Design and ship customer-specific AI configurations: Playbook automations, KB ingestion and chunking strategies, triage and tagging taxonomies, custom integrations against the customers stack via APIs and webhooks.
• Own deployments end-to-end. From discovery through go-live through stabilisation.
• When the customers automation misfires on a Tuesday morning, youre the one who diagnoses it, fixes it, and explains what happened in language their ops lead understands.
• Make AI trustworthy account-by-account. Build per-account golden datasets, run evals against the customers real traffic patterns, and gate rollouts on measured quality — the same eval-first discipline we apply to products, applied to deployments.
• Be the products field intelligence. Every gap you hand-build is a roadmap signal.
• Youll work closely with the AI product and engineering teams to turn repeated custom work into product capabilities, and youll write up deployment patterns so the next account is faster than the last.
• Drive adoption, not just go-live. Partner with CSMs on activation: train champion users, instrument usage, and iterate until the customers team reaches for the AI features by default.

Nice to have

• Experience at a B2B SaaS company in a solutions, implementation, or platform engineering capacity.
• Familiarity with customer support / CX tooling (helpdesks, shared inboxes, ticketing systems).
• Exposure to LLM observability and eval tooling (Langfuse, LLM-as-judge patterns, golden datasets).
• Prior experience as a founder, early-stage employee, or consultant — anything that taught you to scope ambiguous problems with limited support.

What this role is not

• Not a sales engineer role. You enter after the deal (or late in it, for technical validation on strategic accounts). You dont carry a quota.
• Not a support escalation role. You build; you dont run a ticket queue.
• Not a travel role. Embedding happens through the customers Slack, their Hiver workspace, and recurring working sessions.

Why this role matters
Hiver is moving from AI features to AI outcomes. The companies winning in AI-first SaaS — from Palantir to OpenAI to Databricks — learnt the same lesson: the hardest part of AI isnt the model, its making it work inside a real organisations messy reality. This role is how we do that for our most important customers, and how what we learn there shapes the product for everyone else.