Responsibilities
About the role
While you re reading this, developers and enterprises across India are wiring Gnani s speech APIs into their products - running ASR over noisy phone lines and generating TTS in a dozen Indian languages, in real time and at scale. You re the person who makes that integration actually work: the first responder when accuracy dips, the engineer who turns a prospect s can it handle our audio into a working benchmark, and the technical voice the developer community hears back from.
As an AI Solutions Engineer on our Speech API team, you sit between our ASR/TTS platform and the people building on it. You own the customer s technical journey end-to-end - onboarding, accuracy validation, debugging, custom pipelines, and production quality - and you feed what you learn in the field straight back to the teams building the models. This is a hands-on, customer-facing engineering role for someone early in their career who is technically sharp, obsessed with the customer, and communicates like a professional.
Core mandate
Customer success onboarding
Get prospects and customers live on the ASR/TTS APIs fast, and keep them successful
Solutions custom engineering
Design custom speech pipelines and integrations that fit enterprise needs
Benchmarking quality
Run customer and internal benchmarks; spot-check production quality on a schedule
Developer advocacy
Be the technical face for the developer community and turn field signal into product feedback
What you ll drive
Customer onboarding support
- Own technical onboarding for prospects and customers - API integration, authentication, streaming vs. batch, sample code, and first-call success
- Be the first responder on ASR/TTS issues - triage, reproduce, root-cause, and resolve integration and accuracy problems
- Run accuracy checks with customers - set up test sets, interpret WER/CER for ASR and naturalness/MOS for TTS, and explain what the numbers actually mean
- Hand-hold prospects through evaluation so they can reach a confident buying decision
Solutions custom engineering
- Design and build custom speech solutions for enterprise clients - ASR/TTS configuration, custom vocabulary and pronunciation, and end-to-end pipelines
- Translate ambiguous customer requirements into concrete, working technical solutions and reference implementations
- Build demos, POCs, and sample integrations that show the platform at its best
- Partner with the ASR/TTS research and platform teams to shape features around real customer needs
Benchmarking quality
- Run structured benchmarks - internal and customer-supplied - across languages, accents, domains, and audio channels
- Design fair, honest evaluations: representative test sets, apples-to-apples comparison, and no overfitting to a single benchmark
- Perform periodic spot-checks on production quality and flag regressions before customers do
- Turn recurring field failures into clear, reproducible reports the model teams can act on
Developer experience community
- Be the technical, front-facing voice for the developer community - docs, forums, sample code, and integration guides
- Gather developer feedback and channel it into the product roadmap
- Represent Gnani s speech APIs in developer conversations and make the integration experience genuinely good to build on
Who you are
EXPERIENCE WE RE LOOKING FOR
- 2-3 years building speech / ML products, integrating APIs, or in a customer-facing technical role (solutions, support, or field engineering)
- Hands-on experience in at least one of: building an ASR, TTS, or speech-related product; handling technical customers; or building custom solutions and integrations
- A track record of owning technical problems end-to-end and seeing them through to a working outcome
- Strong work ethic, high ownership, and comfort in a fast-moving environment
- Customer-obsessed, with clear and professional written and verbal communication
WHAT MAKES A STANDOUT CANDIDATE
- Hands-on with speech APIs - streaming and batch ASR, TTS, and real-time audio integration
- Comfortable reading and writing integration code - Python plus at least one of REST / gRPC / WebSocket clients
- Familiarity with speech evaluation - WER / CER for ASR, naturalness / MOS for TTS, latency and RTF - and what they mean for a customer
- Understanding of audio fundamentals - sample rates, encodings, codecs, and telephony / contact-centre audio
- Exposure to the Indian voice landscape - multilingual, code-switched, and accented speech
- Experience writing developer-facing docs, guides, or sample code
TECHNICAL FLUENCY - GOOD TO HAVE
- Python for scripting demos, POCs, evaluation harnesses, and debugging
- API integration patterns - auth, streaming, retries, batching, and error handling
- Basic deployment