The Senior AI Developer/Full Stack is a hands-on role within the SalesSuperStar AI team, focused on building the core systems that power our AI-driven Custom Agent and integrated Light CRM platform. This role is critical to the development of a fast-moving, mission-critical MVP, ensuring that frontend, backend, integrations, and platform workflows come together in a scalable, reliable, and production-ready environment.

Overview

We are building a production AI outbound sales platform that combines a real-time AI Custom Voice Agent with an integrated Light CRM and operator platform. This is a hands-on senior role for someone who can design, build, and productionize AI systems while also contributing across the full stack.

The role spans real-time voice workflows, LLM serving, RAG, evaluation, fine-tuning, backend services, and the frontend tools needed to operate and monitor the platform. The environment includes FastAPI, Python, React, TypeScript, PostgreSQL, Redis, Kafka, Twilio-based telephony, local model serving, and AI evaluation tooling.

Position Summary

We are looking for a senior developer with strong AI engineering depth and solid full-stack capability. You will help build and improve a real-time AI voice system covering telephony, STT, TTS, LLM inference, retrieval, prompt orchestration, evaluation, and production deployment, while also supporting the surrounding SaaS platform.

This role goes well beyond simple API integration. You will work across a high-concurrency production architecture that combines CRM integration and the AI Custom Agent, including FastAPI, Python, React, PostgreSQL, Redis, Kafka, telephony, voice AI, vector retrieval, and observability. The right person can translate AI concepts into reliable, measurable business systems.

Key Responsibilities

AI Agent Development

• Design, build, and improve the core AI Custom Agent used in a live outbound sales environment.
• Develop and optimize prompt orchestration, context management, memory handling, tool calling, and agent workflows.
• Build and maintain RAG pipelines using modern retrieval frameworks, vector databases, embeddings, and conversation state persistence.
• Improve agent quality through structured evaluation, benchmark testing, error analysis, and systematic iteration.
• Work on voice AI workflows spanning telephony, audio streaming, STT, TTS, turn-taking, and response generation.

LLM Inference, Fine-Tuning, and Optimization

• Deploy, run, and optimize open-weight models in production using inference frameworks such as vLLM.
• Work directly on throughput, concurrency, latency, and time-to-first-token optimization for real-time AI workloads.
• Build and maintain fine-tuning workflows using LoRA or QLoRA and the Hugging Face ecosystem.
• Support training, evaluation, quantization, checkpoint management, and deployment workflows for domain-specific model improvements.
• Help determine when prompting, RAG, fine-tuning, or workflow design is the right solution to a given business problem.

Backend and Platform Engineering

• Build and maintain backend services in Python and FastAPI that power the AI management layer, orchestration logic, APIs, and platform integrations.
• Design clean internal and external APIs, webhooks, and service boundaries across CRM, telephony, AI, and automation layers.
• Implement robust data models and backend workflows using PostgreSQL, Redis, Kafka, and related services.
• Ensure fault tolerance, observability, and production reliability across distributed services and background processes.

Full Stack Product Development

• Contribute to the operator and admin platform using React and TypeScript.
• Build interfaces for lead workflows, call monitoring, agent controls, prompt management, quality review, analytics, and platform operations.
• Partner with product, QA, and operations stakeholders to ensure the platform is usable, stable, and aligned to business workflows.

Production Engineering and Collaboration

• Work closely with infrastructure, telephony, QA, and product teams to move AI capabilities into stable production.
• Contribute to CI/CD, testing, observability, security, and engineering best practices.
• Participate in architecture reviews, code reviews, and technical decision-making.
• Help establish standards for AI engineering, experimentation, deployment, and platform reliability.

Key Qualifications, All Must-Haves

• 7+ years of software development experience, including significant hands-on backend engineering.
• Strong backend Python capability and proven production experience with FastAPI.
• Strong full-stack development experience, including React and modern frontend application development.

Hands-on experience:

• Building AI applications beyond simple API wrappers, including prompt orchestration, context handling, memory, and multi-step workflows.
• Designing and implementing RAG architectures, including ingestion, chunking, embeddings, retrieval strategy, vector stores, and context injection.
• Deploying and running open-weight LLMs in production, ideally using vLLM.
• With fine-tuning methods such as LoRA or QLoRA, ideally using tools in the Hugging Face ecosystem.
• With evaluation frameworks and systematic improvement of LLM outputs, including prompt iteration, response quality assessment, benchmark-based testing, and structured evaluation workflows.
• With real-time AI systems or voice workflows, including telephony, STT, TTS, streaming, or other low-latency interactive architectures.
• Designing and building production APIs, distributed services, and integration-heavy systems.
• With PostgreSQL and Redis, including state handling, caching, queues, and pub/sub patterns.
• Integrating third-party SaaS and platform APIs, including CRM, telephony, or workflow automation systems.
• With Docker-based environments and production deployment workflows.
• With monitoring, logging, observability, and troubleshooting in live production environments.

Preferred Additional Strengths

Experience as follows:

• LangChain, LangGraph, Qdrant, sentence-transformer pipelines, or similar orchestration and retrieval tooling.
• DeepEval, lm-eval-harness, Langfuse, MLflow, Locust, or similar evaluation and benchmarking tools.
• Pipecat, Twilio, Deepgram, ElevenLabs, or related voice AI stack components.
• Optimizing high-concurrency inference workloads in production.
• On-prem or private AI infrastructure.
• Building high-throughput sales, contact center, or communications platforms.

Closing Profile

This role is for a builder who is genuinely AI-first, not a traditional full-stack developer who has only lightly touched AI. You should be comfortable owning the hard parts of applied AI engineering in production, while also being capable of contributing across the supporting full-stack platform. The right person will help us accelerate model quality, voice workflow reliability, production performance, and overall, AI platform maturity.

Diversity & Accessibility

At OWE, we believe that diverse perspectives fuel innovation. We are committed to fostering an inclusive environment that values equity and accessibility. If you require accommodations at any stage of the hiring process, please contact us at recruitment@ontariowholesaleenergy.com — we’ll be happy to assist.