AI/ML Engineer

Benefits

• 401(k) matching
• Health insurance
• Opportunity for advancement
• Paid time off
• Training & development

Essential Job Responsibilities

• Implement AI workflows, including workflow logic, intent classification, entity extraction, and multi-step decision processes
• Build and tune LLM-powered features using OpenAI, Anthropic, and other APIs — including prompt engineering, structured output parsing, and function calling
• Develop NLP/classification systems for routing
• Design and run model evaluations: build test sets, measure accuracy/latency/cost, iterate on prompts and model choices
• Tune confidence thresholds (e.g., 90%/70% auto-vs-human routing) and monitor drift in production
• Integrate AI and workflows with existing enterprise systems
• Implement observability for AI systems, including structured logging, prompt and pipeline tracing, and cost attribution per execution
• Collaborate with cross-functional teams to operationalize and iterate on prompts
• Contribute to code review, technical design docs, and on-call rotations
• Delivers responsive and professional customer support by diagnosing and resolving technical issues, communicating clearly with non-technical users, and ensuring timely follow-up to maintain high levels of customer satisfaction
• Keep up with rapidly evolving technologies, algorithms, and industry trends
• Work closely with business managers and other engineers to align AI strategies with company goals

Required Qualifications

• Strong experience and knowledge in Python and TypeScript
• Hands-on LLM API experience: Anthropic Claude, OpenAI, at least one other provider; function calling, JSON mode, streaming
• Prompt engineering as a systematic practice — not ad-hoc wordsmithing, but versioned prompts with test sets and regression checks
• NLP fundamentals: intent classification, NER, embeddings, semantic search (pgvector, Pinecone)
• Create evaluation approaches that define meaningful success metrics, build test datasets, and track performance against client-specific KPIs.
• Model selection tradeoffs: when to use GPT-5 vs. Claude Sonnet vs. Haiku, when to fine-tune, when RAG is the right pattern
• Voice AI STT/TTS integration: Deepgram Nova/Aura, Whisper, ElevenLabs
• Real-time streaming audio handling — latency budgets, interruption handling, turn-taking
• Twilio Voice API, SIP trunking concepts, WebRTC awareness
• Software Engineering Backend development: REST APIs, async patterns, message queues
• Databases: SQL (PostgreSQL, SQL Server), basic schema design; vector DB basics
• Cloud: Azure (AWS/GCP acceptable), containerization (Docker)
• Git, CI/CD, code review fundamentals

Preferred Skills

• MLOps tooling: Langfuse, MLflow, Weights & Biases
• Familiarity with model adaptation techniques, such as fine-tuning (e.g., LoRA, supervised approaches)
• Experience with frameworks: LangChain, LlamaIndex, Pydantic AI — used judiciously, not as a crutch