Company Description

MagickVoice is transforming how businesses communicate by delivering intelligent, AI-powered voice solutions for customer engagement, internal communications, and voice-enabled applications. The platform combines advanced artificial intelligence with cutting-edge voice technology to create smart, scalable, secure, and seamlessly integrated communication experiences. Organizations use MagickVoice to handle millions of interactions across channels while maintaining enterprise-grade security and compliance.

With multilingual support, real-time analytics, and developer-friendly APIs, MagickVoice helps teams enhance customer experiences and improve operational efficiency. The company is committed to democratizing intelligent voice communication and is trusted by industry leaders who value continuous innovation in AI-powered voice.

What you'll actually do:

Deploy voice bots for customers, end to end. Sit with a customer, understand what the call is supposed to achieve, and build it — prompt template, opening message, variables, guidelines, escalation triggers, analytics dimensions. Configure it across languages. Test it by calling yourself, then by calling ten people who aren't you.

Run and tune campaigns. Upload the contact list, set the schedule, watch the first fifty calls land. Read transcripts. Find the two lines in the system prompt causing 30% of drop-offs. Fix them. Rerun. This loop is most of the job and it's the part that decides whether a customer renews.

Know the product better than anyone. Every feature, every setting, every edge in the pipeline — STT, TTS, latency, interruption handling, retries, webhooks, the API, the dashboard. When a customer asks "can it do X," you should know the answer without checking, and know the workaround when the answer is no.

Write the playbook.



We're building the same kinds of calls repeatedly — admissions follow-up, fee reminder, event confirmation, lead qualification, feedback collection. Turn what you learn in the field into documented, reusable scenarios so the next deployment takes two days instead of two weeks. If it lives only in your head, it doesn't count.

Break things before customers do. You'll be first in line on recent features — testing them against real deployment scenarios and telling the team plainly what's broken, what's confusing, and what nobody will use. Your bug reports carry weight because you have call recordings to back them up.

Bring the field back to the product team. You'll be the loudest signal we have on what to build next. We expect you to argue for it.

What you'll actually do:

- 2+ years building or deploying software. Backgrounds that work: backend or full-stack engineering, solutions/implementation engineering, technical customer success with real hands on code, or a founding engineer at something small.
- Comfortable in Python or JavaScript — enough to read our codebase, script a data cleanup, hit an API with curl, and debug a failing webhook without waiting for someone else.
- You write clearly. Prompt engineering is writing. Documentation is writing. Explaining to a customer why their call flow won't work is writing. This matters more here than most engineering roles.
- You're comfortable talking to customers directly — including when something is broken and they're unhappy about it.



You can hold that conversation without disappearing or over-promising.
- You use AI tools daily and fluidly. Claude, GPT, Gemini, whatever ships next month — you should have opinions on which one to reach for and be happy to switch when a better one lands. We're model-agnostic by design and we want people who are too.
- You like ambiguity more than process. No one will hand you a ticket queue. You'll figure out what's most broken and go fix it.
- Indian language fluency is a real advantage — Telugu, Hindi, Tamil, Kannada or Marathi. A lot of our quality work is listening to calls and knowing when the bot sounds off.

What you'll actually do: We're an early-stage team, and we'd rather set expectations now than in month three.

Your scope will shift. Some weeks are customer deployments, some are internal tooling, some are spent deep in transcripts improving a single prompt. Variety is the norm here, not the exception.

Campaigns run on customers' schedules. Calls go out when people are likely to pick up, which occasionally means a launch window outside standard hours. We plan these in advance, rotate them across the team, and make sure the time comes back to you.

Some of what you'll document doesn't exist yet. You'll be defining the playbook as you build it rather than inheriting one — which is a genuine opportunity to shape how we work.

Ownership comes early. You'll run deployments end to end with real decision-making authority, and with the team behind you when something breaks.

Occasional travel to customer sites, mostly within India.

If you're looking for a clearly defined ladder and a narrow specialisation, we're honest that we're too early to offer that today. If you'd like to see your work land on real phone calls within the week, you'll enjoy it here.

📌 Forward Deployed Engineer (Hyderabad)
🏢 MagickVoice
📍 Hyderabad