AI Evangelist - Architect Advanced

Role: AI Evangelist - Architect Advanced

Location:
San Leandro, CA (onsite, locals only)

Duration: 6 months

Pay Rate: $65 W2 and $80-85 C2C

Annual Salary: $180K per annum

We’re seeking an inquisitive, hands‑on AI Evangelist who can turn ideas into shipped capabilities—applying generative and agentic AI to enhance existing products and build new tools. You will partner with product, engineering, data, and business stakeholders to client high‑impact use cases, build working prototypes, and guide production adoption, while championing responsible‑AI practices and measurable outcomes.

• Educate & influence:
  Lead demos, brown‑bags, and workshops to raise AI fluency across product, engineering, and business teams; translate complex AI concepts into clear, outcome‑oriented narratives.
• Client value:
  Run structured discovery (problem framing, ROI/feasibility) to identify high‑leverage AI use cases in current applications and greenfield tools.
• Prototype fast:
  Build end‑to‑end proofs of concept (POCs) using LLMs and agent frameworks, moving from idea → prototype in weeks, not months.
• Integrate & ship:
  Partner with product and platform teams to embed AI features into existing stacks (APIs/services, front‑end surfaces, workflows), hardening POCs for production.
• Agentic systems:
  Design agent workflows (planning, tool‑use, retrieval, guardrails) for tasks like intelligent assistance, automation, and decision support.
• Architecture & ops:
  Define reference architectures for RAG, tools/plugins, orchestration, observability, evaluation, and cost/performance tuning.
• Governance:
  Embed Responsible AI (safety, privacy, security, compliance), data governance, and evaluation frameworks (offline/online) into delivery.
• Measurement:
  Establish success metrics (quality, latency, adoption, cost per task, deflection, NPS/CSAT) and run experiments/A‑B tests to validate impact.
• Partner ecosystem:
  Evaluate vendors and open‑source components; guide build‑vs‑buy decisions; contribute reusable assets and playbooks.
• Champion change:
  Remove adoption blockers, capture learnings, and scale wins via internal communities, templates, and enablement content.

• Total 15+ years of experience in Software engineering, with 8+ years in ML engineering (or equivalent) with 2+ years delivering generative AI features or platforms end‑to‑end.
• Demonstrated ability to prototype and code: one or more of Python/Type Script/Java, plus modern API and microservice patterns.
• Hands‑on with LLMs and agentic patterns: prompt engineering, RAG, tool‑calling/function‑calling, agents/planners, evaluation.
• Experience with at least one cloud (Azure OpenAI, AWS Bedrock, Google Vertex AI) and vector/search stacks (Pinecone, FAISS, Elasticsearch/Open Search, pgvector).
• Familiarity with Lang Chain/Lang Graph, Llama Index, OpenAI/Claude APIs, and model hosting (managed endpoints or self‑hosted).
• Solid understanding of security, privacy, governance, PII handling, prompt‑injection mitigation, abuse monitoring, and auditability.
• Interpersonal excellence: persuasive communicator and facilitator; comfortable with exec briefings and hands‑on pairing with engineers.
• Strong product sensibilities: able to frame problems, define success metrics, and iterate with user feedback.

• Experience operationalizing AI features: eval harnesses (LLM‑as‑judge/human‑in‑the‑loop), observability (trace logs, prompt/versioning), and cost/perf tuning.
• Background in MLOps (feature stores, CI/CD for ML, model/version management) or platform engineering for AI services.
• Domain experience in regulated industries (e.g., financial services, healthcare) and threat‑modeling for AI systems.
• Contributions to OSS, internal frameworks, or thought leadership (blogs, talks, playbooks).

• 3–5 shipped AI capabilities improving core KPIs (quality, cycle time, cost per task, or revenue uplift).
• Reusable assets: reference architectures, starter repos, guardrail/eval templates, and adoption playbooks.
• Organization enablement: > 200 employees enabled via workshops/office hours and a sustained…