19 Sep
|
InstaBus
|
India
About InstaBus
India's private intercity bus industry moves millions of passengers a day and still runs on phone calls, WhatsApp forwards and spreadsheets. We build AI products that change that.
Our first product, InstaVoice, is a 24×7 AI voice agent that speaks natural Hinglish and handles the entire passenger journey — booking confirmations, boarding point coordination, delay updates, feedback capture. It is live with real operators, on real routes, talking to real passengers every night.
We are a small team. What you own actually ships.
The role:
You are the human in the loop for an AI system that talks to thousands of passengers.
Two things depend on you: the data our AI agent speaks from, and the judgement call on whether it spoke well. Both are the difference between an AI product that operators trust and one they switch off.
This is not a ticket queue. You will sit between our operators, our AI and our product team, and you will be the first person to know when something is wrong.
What you'll own
1. AI voice quality assurance
- Listen to live AI call recordings daily and assess them against a quality rubric — comprehension, response accuracy, tone, interruption handling, call resolution
- Identify failure modes: misheard place names, wrong language switches, dropped context mid-call, awkward phrasing, voice quality issues
- Log, categorise and prioritise issues so the product team knows what to fix first
- Track whether fixes actually worked, call over call
- Flag anything that affects a passenger's real journey — immediately
Your notes directly shape what our engineers change in the agent. You are the closest thing we have to the passenger's ear.
2. Passenger data operations
- Pull daily passenger manifests from operator GDS/back-office systems across multiple bus operators and routes
- Use AI tooling (Claude) to clean, validate and standardise those manifests into our platform format — deduplication, phone merging, name normalisation, boarding/drop point mapping, agent vs. OTA classification
- Cross-check every file against source totals and flag discrepancies before they reach the dashboard
- Push validated files to the InstaBus dashboard on schedule, including evening and night departure cycles
- Spot patterns in bad data and tell us how to fix them at source — most of our processing rules came from someone noticing something twice
You will not be doing this by hand. You'll be operating an AI-assisted pipeline and being accountable for its output. Learning to prompt, correct and QA an AI system is a large part of the job.
Who we're looking for
- Fluent in Hindi and English and Southern Regional Languages (esp. Kannada), comfortable with Hinglish as it is actually spoken — you'll be listening to calls, not reading transcripts
- Solid attention to detail. You notice when a number doesn't add up and you don't let it slide
- Comfortable with Excel/Google Sheets and willing to learn AI tools properly (we'll train you; curiosity matters more than existing skill)
- Able to write clearly — your issue logs are a product input, not a formality
- Self-managing. We don't count hours — we count whether the file went up before the bus left. Buses run on schedule, so our data has to
- Any graduate. Freshers welcome
Bonus: exposure to travel/transport/logistics operations, customer support quality monitoring, or any hands-on use of ChatGPT/Claude in real work.
How we work
Hours and days are flexible. We don't run a fixed shift. What we need is roughly 8 hours of real work a day and coverage of the departure cycles — manifests are pulled and uploaded through the morning, evening and late-night slots, and calls are reviewed daily. You'll plan your own week in coordination with the rest of the team and our interns, so that someone is always covering the cycles. If you want to take a Wednesday off and work Saturday, that's a conversation, not a problem.
Confidentiality is non-negotiable. You will handle real passenger names, phone numbers and travel details. That data stays inside our systems, always. We'll ask you to sign an NDA.
Why this is worth your time
- You'll learn AI operations from the inside — prompting, evaluation, quality rubrics, failure analysis. This is a fast-growing skill set and very few people have real production experience in it
- Direct access to the founder and the product team. No layers
- Clear growth path into AI Operations Lead, Customer Success, or Product Operations as we scale across operators
- Early-stage startup with paying customers and real deployments
Pay: From ₹25,000.00 per month
Benefits:
- Flexible schedule
Work Location: In person
📌 AI Operations Associate — Voice Quality & Data (India)
🏢 InstaBus
📍 India