AI Development Lead

Responsibilities

• Own technical architecture for AI microservices, data flows, and LLM prompts/guardrails; enforce CI/CD and observability.
• Lead model development: define strategies for LLM tuning (instruction-tuning, LoRA/PEFT, adapters), data curation/synthetic data, hyperparameter search, and offline/online evaluation (golden sets, A/B, canary).
• Build NLP capabilities: intent/slot models, NER, sentiment, summarization, classification/ranking; select architectures (Transformers) and optimize with metrics like F1, precision/recall, ROUGE, BLEU, and embedding similarity.
• Stand up AIOps/MLOps (monitoring, alerting, fallbacks, retraining), define accuracy/SLA benchmarks, and guide QA.
• Lead voice interfaces (STT/TTS), telephony, and real-time dialogue management across channels.
• Oversee CRM, calendar, and payment integrations; ensure conflict resolution, secure transactions, and compliance (GDPR/CCPA/PIPEDA).
• Drive experimentation for model/STT/TTS selection and orchestration; document outcomes, mentor engineers, and produce partner docs/SDKs.

What You’ll Lead:

• Front Desk Agent (B2C): Microservices + CI/CD; LLMs (Gemini/GPT) with fine-tuned or adapter-based variants; LiveKit/Twilio voice; personalization/analytics; production hardening. Integrate WellnessLiving booking and Nuvei/Paragon payments; robust call/session management.
• Business Coach (B2B): Coaching workflows powered by NLP/NLU; CRM integrations (Salesforce/HubSpot/Zapier); KPI-aware guidance; live voice summaries.
• Sales & Marketing Suite: Workflow engine; omnichannel messaging (SendGrid, Twilio, WhatsApp); LLM content; pricing/monetization; dashboards; Google/Meta Ads.
• Omnichannel Message Center: Orchestration (Kafka/SQS); Twilio/SendGrid/FCM/SES; sentiment-based routing/escalation; encryption/audit.
• Scale & GTM: Monitoring and performance/latency improvements; UAT and staged rollout; partner SDK support and model rollout governance.

Minimum Qualifications:

• 7+ years software/AI engineering, including 3+ years leading AI product teams (LLMs, conversational AI, or voice).
• Hands-on LLM tuning: instruction-tuning and/or LoRA/PEFT, prompt design and evaluation, structured tool/function calling, RAG system design (chunking, embeddings, re-ranking).
• Solid NLP grounding: text preprocessing, tokenization, embeddings, classification/NER/summarization/semantic search, and evaluation (F1/ROC-AUC/ROUGE/BLEU).
• Delivered production AI with LLMs, telephony, and third-party APIs (CRM/payments/messaging).
• Hands-on cloud microservices & data (AWS, Aurora Serverless, Redshift), Docker, CI/CD; testing and performance tuning.
• Governance/compliance experience (GDPR/CCPA/PIPEDA); proven path from prototype to production with reliable ops.

Nice to Have:

• Quantization/distillation, retrieval re-ranking (e.g., cross-encoders), and vector DBs (FAISS/pgvector/Pinecone).
• LiveKit/Twilio/SIP; multilingual NLP; time-series and propensity/lead-scoring models.
• Kafka/SQS; multi-channel orchestration (SMS/email/push/WhatsApp); real-time voice UX.

How We’ll Measure Success:

• Front Desk Agent: 70–85% automation of booking in year one; sub-2s response under load.
• Business Coach: >90% intent recognition.
• Message Center: 85–95% classification accuracy, >90% sentiment precision, ~10s first response.
• Sales & Marketing: >85% lead-scoring accuracy; 10k+ workflow triggers/hour with <1s response.
• Continuous latency/error-rate reductions via tuning and benchmarking.

Please note that only those selected for an interview will be contacted.

We appreciate you taking the time and look forward to reviewing your application!

WellnessLiving is proud to be an equal opportunity employer. We base employment decisions solely on qualifications, experience, and business needs. We do not tolerate discrimination or harassment of any kind. All qualified applicants will receive consideration without regard to race, color, religion, creed, gender, gender identity or expression, sexual orientation, national origin, disability, age, genetic information, veteran status, marital or family status, or any other status protected by applicable laws.

We utilize AI to generate summaries of interview notes as part of our candidate evaluation process. This helps ensure a fair and consistent review while maintaining a human-centered hiring approach.