Forward Deployed Agentic Engineer
- Hiring from
- South Africa
- Work type
- Hybrid
- Posted
625,597 remote jobs, straight from company career pages
100% free · New jobs every minute
Show job descriptionHide job description
Forward Deployed Agentic Engineer, BlueSky Digital Solutions
South Africa, hybrid from Cape Town or Johannesburg
We're hiring an engineer who builds with Claude every day and wants that work to reach production inside a large enterprise's own systems, under the client's own review, which very few engineering jobs in South Africa offer right now. BlueSky is an Anthropic partner, a Salesforce Summit Partner and an AWS Advanced Consulting Partner, and we run multi-year AI, AWS and Salesforce delivery for large South African enterprises, from banks to telecoms and retail.
The job is part build and part enablement, advocacy and thought leadership. On the build side you'll sit with one of our architects on a live client programme, getting the client's backend services ready for agents to use safely, which means things like a payment that can only be released against the reference it belongs to, evidence that stays bound to its own case, retries that never pay twice, and an action the agent proposes that a person confirms before anything moves. You'll also help harden the pipeline the client's own engineers use with Claude Code: sandboxing, hooks that stop the dangerous commands, and a security review on every change that runs on its own. Beyond the build, you'll help the client's engineers work this way themselves, make the case for safe agents with their architects and risk teams, and share what we learn through talks and writing.
What we're looking for
- Three or more years building and shipping software, with real time spent on backend services.
- You've built real things with Claude Code or the Claude API, with MCP servers, agent loops and evals, and you can show us the work and tell us what you check before you trust what an agent hands back.
- You're comfortable in backend services, Java or something close to it, because that's where most of the agent-ready changes land.
- You know AWS or Salesforce well enough to find your way around a client's environment. You don't need both.
- You think about how an agent gets misused, not just how it works: untrusted input, sensitive data and the power to act, and what happens when all three meet.
- You treat security and governance as part of the build, not a gate at the end: least-privilege access, audit logs, secrets handling and data protection rules like POPIA, and you're comfortable walking a client's security team through why your design is safe.
- You write clearly. Answer first, plain words, and you say so when you're not sure.
What you'll do
- Build and ship agent-ready changes and agent guardrails inside the client's environment, through their own change control.
- Write the tests and evals that prove it behaves, including the ones that try to break it.
- Build governance and security in from the start, so the agent gets only the access its job needs, every action it takes leaves an audit trail a risk team can follow, client data stays where POPIA and the client's own policies say it must, and the client's security and risk teams have the evidence they need to sign it off.
- Own pieces of the programme end to end, from the brief through to the client's sign-off, working alongside the architect you sit with.
Improve the shared plugins our skills and agents ship in, so when one gets something wrong on a job you fix it for the whole team with a pull request that carries its tests and registry entry, rather than patching your own copy.
Who you are
The tools change every few weeks, so we care more about how you think than about any one framework. We're looking for someone who owns what they ship, who can make the call on a technical approach and explain it to a client's own architects, tries the angle nobody else did, and still says so when the plain option is the right one.
That's our four values in practice, Embrace Difficult, Make It Better, Pay It Forward and In It Together, and the last one is why every piece of work gets a second pair of eyes before it reaches a client, yours included.
Where this goes
By month six we'd expect you to run one workstream end to end without someone checking each step, and by month twelve to lead client conversations on your own, with the Agentic Solution Architect role as the next step for people who want it. If you've never worked in banking, telecoms or retail, that's fine, as long as you're keen to learn how a large enterprise thinks about risk, because that's the part we'll teach you.
What you get
Your own Claude seat from day one and time to experiment with it, Claude Certified Developer Foundations paid for by us, and the AWS or Salesforce certification you want next. You also get 20 days annual leave.
How we hire
A short intro call, then a session where you walk us through something you've built with an agent and what you'd change, and we aim to give you an answer within two weeks of that session. Claude helps us sort applications by role, and a person reads and decides on everyone. We keep your details for 12 months only to run this search, so tell us if you'd rather we didn't.
The rest
We work hybrid, three days a week either at our office or at the client and the rest remote, and you'll get technical training, leadership development and a mentor who'll guide you, challenge you and back you. Depending on the client, expect background checks before you start. If you'd like to see how we work before you apply, have a look at our culture page at https://bsky.co.za/culture/.
To apply, go through the posting on our careers page, which asks three short questions: something you've built with an AI agent recently, what you do when an agent hands you work you think is wrong, and whether you'd rather build, run or lead.
