About us
Techinnov (Creative Techinnov Software Solutions Inc.) is a Montreal-based software company. For seventeen years we have been the technology partner behind a federal-scale language training platform used every day by thousands of learners. We build production software, for the training industry.
About the role
We are building an AI conversation partner: a voice-first tool that lets learners practice spoken conversation against structured pedagogical content, integrated into an existing learner portal. The stack is Azure-centric with intentional model portability across Azure OpenAI, Google Gemini, and specialist voice platforms.
You will join a small engineering team as a full-time contributor from day one. This is a hands-on build role. If the last twelve months of your work does include shipped, running code that other people rely on, this is the right role for you.
What you will do
- Design and build the voice AI pipeline end-to-end: cascaded (STT + LLM + TTS) and/or native realtime
- Integrate LLM inference with the portal and learner-state layer
- Implement pedagogical calibration in the prompt layer, co-developed with the client's Innovation team
- Stand up an automated evaluation harness: rubric-scoring prompts, held-out test sets, quality gates
- Contribute to vendor benchmarking and architecture decisions
- Write code that the existing team can maintain after your contract ends
Must have
- 5+ years shipping production software in a modern stack (Python, TypeScript, or equivalent)
- 2+ years hands-on with LLM APIs (OpenAI, Anthropic, Google, or equivalent) in production, not prototypes
- Direct experience with voice AI: at least one of STT/TTS pipelines, native realtime voice APIs (Gemini Live, Azure Realtime, OpenAI Realtime), or managed voice platforms (ElevenLabs Agents, LiveKit, Vapi)
- Working knowledge of prompt engineering and RAG patterns
- Cloud experience, Azure preferred (GCP or AWS acceptable)
- Strong written and verbal English
- Available to start the week of September 15, 2026 for onboarding
Nice to have
- Prior work on educational technology, particularly language learning
- French language proficiency
- Familiarity with Canadian federal data residency and compliance requirements
- Experience integrating with enterprise portals, SSO, feature-flag systems
- Track record building evaluation frameworks for LLM applications
This role is not for
- AI strategy consultants
- ML researchers without production shipping experience
- Prompt engineers whose portfolio is UI-layer only
- Anyone whose recent work is architecture diagrams and not code
How to apply
Send a short cover note and your resume, GitHub, or portfolio
Please include:
- One voice-AI or LLM project you built end-to-end, in one paragraph
- The stack you used and what you specifically built (versus what you specified for someone else to build)
- Whether the project is still in production and how it is used
- Your available start date and expected day rate
Applications reviewed on a rolling basis.
Pay: $130,153.78-$198,140.33 per year
Work Location: Remote
