Founding AI Engineer (LLMs & Agents)
Rethink recruitWhile most AI companies focus on digital industries, Skiffra designs AI-native orchestration systems for the sectors that extract, move, and build the ecosystems around us. We turn messy, high-stakes operations into intelligent, adaptive workflows that improve decisions in real time.
Our first proving ground is mining and natural resources, but the platform is modular by design and built to scale into any industry where decisions are expensive and consequences are real.
We’re looking for a Founding AI Engineer (LLMs & Agents) to architect and build the core intelligence behind Skiffra. This is a founding-level role for a high-leverage builder who can move fluidly between prompt engineering, backend systems, and model optimization—turning agentic workflows from whiteboard ideas into production-grade systems handling real enterprise data at scale.
What You’ll Do
Architect & Build Agentic Systems
-
Design, prototype, and harden production-grade orchestration frameworks
-
Leverage LLMs, persistent memory, and tool integrations to navigate complex, non-linear enterprise workflows
-
Move agentic systems from experimentation to reliable, enterprise-ready infrastructure
Build the Data Flywheel
-
Implement sophisticated RAG (Retrieval-Augmented Generation) pipelines
-
Ensure high-fidelity context retrieval from fragmented, messy operational data sources
-
Optimize context management, embeddings, and retrieval performance
Evaluate & Iterate
-
Develop internal benchmarking and evaluation suites
-
Improve model reliability, grounding, hallucination resistance, and latency
-
Instrument systems for observability and continuous performance improvement
Strategic Modeling Decisions
-
Architect intelligence stacks with pragmatic trade-offs between:
-
Frontier reasoning models (e.g., GPT-5, Claude 4 Opus) for long-horizon planning
-
Fine-tuned open-weight models (e.g., Llama, Mistral) for sovereignty and low-latency inference
-
Deterministic heuristics for reliable, cost-efficient guardrails
-
Full-Stack Ownership
-
Work directly with founders to translate ambiguous operational “messes” into clean, scalable AI systems
-
Contribute across backend architecture, infrastructure, and product logic
What We’re Looking For
-
7+ years of systems engineering experience building production-grade software
-
Deep practical experience with modern transformer-based stacks:
-
Embeddings
-
Vector databases
-
Tool-calling frameworks
-
Context window management
-
-
Strong Python engineering skills—you build systems, not just notebooks
-
Reliability-first mindset with strong instincts around failure modes (hallucinations, API timeouts, infinite agent loops)
-
Clear communicator with founder-level ownership mentality
Nice to Have
-
Experience building multi-step workflows with LangGraph, LlamaIndex, or CrewAI
-
Hands-on fine-tuning and PEFT (LoRA/QLoRA) for adapting open-weight models
-
Background in enterprise operations (industrial, logistics, fintech, CRM workflows)
-
Strong, opinionated perspective on where AI creates real value—and where it’s just expensive noise
How We Work
-
Speed over over-planning
-
Simplicity over cleverness
-
Prototypes over slide decks
-
Ownership over handoffs
You’ll work directly with founders, shipping systems that influence real-world operations—not just digital metrics.
Location & Work Model
Hybrid role based in Los Angeles, CA or the San Francisco Bay Area.
Regular in-person collaboration is expected. We are not remote-only.
Work Authorization
Candidates must be authorized to work in the United States.
Visa sponsorship is not available for this role.
Benefits & Compensation
-
Competitive salary and equity as a founding team member
-
Health insurance and retirement plans
-
Eligibility for bonus and incentive programs
-
High-impact role with meaningful ownership in shaping the company’s core intelligence layer
If you care more about shipping real systems than demo-day hype, this role is for you.