About Bolna

Why our Customers love us

1. Ability to train agent and go live with 1000s of calls in less than 2 days
2. Massive increase in ROI - with agents matching human performance at 50% of the cost
3. Effortless up-scale and down-scale based on calling requirements
4. Transparent pricing (only per minute) and best-in-class support

How we work

• High-velocity: Rapid experimentation, lean autonomous teams, and minimal bureaucracy.
• Impact not job titles: We don’t have job titles. Instead, it’s about the impact you have. No task is above or beneath you.
• AI first: We use AI to move faster with higher-quality results. We do this across the whole company - from engineering to growth to operations.
• Excellence everywhere: Everything we do should match the quality of our product.

About the role

What you’ll own

1. The failure ontology. Move Bolna from "call failed" to "call failed because X" - a structured, labelled, growing classification of why Indian enterprise voice agents break. Instrument the platform to auto-tag every call against it.
2. Ground-truth and benchmarks. Partner with 3-5 anchor customers per vertical to link call outcomes to actual business results - did the payment come in, did the candidate progress, did the delivery confirm. Build the "Bolna Collections Benchmark" (and similar) that any new agent can be run against.
3. Production evals and the prompt-diff engine. The in-studio evals are Studio’s; you own the ones that run continuously in production and flag agent degradation. Your pod generates the suggested fixes; Studio owns how they show up in the UI.
4. Experimentation capability. A/B on the platform, model routing randomisers, prompt-variant rollouts. If someone at Bolna has a hypothesis about an agent, you own the infra they use to test it.
5. Model fine-tuning, post-training, token optimisation. Use the corpus we’re accumulating to build vertical-specific models and to drive down inference cost. Co-own the roadmap with ML engineering.
6. The Orchestration/Marketplace layer. Model routing (ASR/LLM/TTS selection by language, latency, cost). Telephony parity across Plivo, Vobiz, Exotel, and IVR. Dialer, concurrency, retry, failover. Latency and reliability budgets.

What you bring

• Technical depth to go head-to-head with ML engineers on eval design, dataset construction, and model routing. You don’t need to write production PyTorch, but you do need to read a research paper and argue about whether its eval methodology applies to our corpus.
• Strong instinct for what a "good" eval looks like vs a Prometheus statistical trap. You can articulate why a completion rate of 95% is not the same as a success rate of 95%.
• Systems thinking. You should be able to trace how a tagging decision in the failure ontology flows through to an agent health score to a prompt diff to a customer email - and what breaks where.
• Comfort with ambiguity at the infra/product seam. Model routing is half product, half devops decision. You shouldn’t need a product management playbook to tell you when to pull rank on which.
• Written clarity. The failure ontology is ultimately a document. The eval methodology is ultimately a document. You should enjoy writing them.

Nice to have

• Background in ML, data engineering, quant, or a previous role building internal eval/experimentation platforms.
• Exposure to telephony, contact centre, or CPaaS infrastructure.
• You’ve shipped something that involved negotiating with a real customer for ground-truth data access.

Success in this role looks like

• 80+% of calls carry structured failure tags within 2 quarters.
• A published, sharable vertical benchmark (e.g. "Bolna Collections Benchmark") that Bolna’s sales team uses in pitches.
• Meaningful improvement in agent success rate after prompt-diff interventions, attributable to the loop you built.
• Model routing that measurably reduces cost per call or p95 latency without quality regression.
• Telephony parity across providers so that no reseller deal is blocked by a CIN compliance issue or billing gap.

What we offer

• Innovative culture: You’ll be part of a generational opportunity to define the trajectory of AI, surrounded by a team pushing the boundaries of what’s possible.
• Growth paths: Joining Bolna means joining a dynamic team with countless opportunities to drive impact - beyond your immediate role and responsibilities.
• Learning & development: Bolna proactively supports professional development, the processes for which are being set up.
• Competitive compensation + meaningful ESOP
• Bengaluru office, in-person team collaboration