Trách nhiệm chính

• Advise and design AI solutions across the full spectrum, from single LLM calls to complex multi‑agent infrastructure, recommending implementation approaches that account for fallbacks, cost/latency budgets, and rate‑limit constraints.
• Architect agentic systems with the right structure for the problem: MCP server design, deterministic workflows, single agents with tool access, orchestrator–worker or A2A multi‑agent topologies.
• Own the evaluation practice that drives technical decisions: eval sets, scoring rubrics, and regression measurement for non‑deterministic outputs, using evidence to settle when to prompt, RAG, fine‑tune, or build an agent.
• Partner with Product from the design phase: consult on where AI genuinely adds value, feasibility, cost, and failure modes, shaping product spec.
• Forecast the direction of AI in the next 2‑3 years and take actions to keep Perform.AI technologically relevant.
• Raise the technical level of engineers across squads: review designs and code, define shared standards, deliver internal AI training programs, and make architecture decisions the whole team builds on.

Yêu cầu

• 8+ years of software engineering experience, with 2+ years building and operating LLM‑based systems in production.
• Hands‑on experience designing and running LLM‑based systems in production, handling fallbacks, cost/latency trade‑offs, rate limits, and structured output reliability at scale.
• Experience building multi‑step agentic systems and choosing the right topology, using frameworks such as LangChain, LangGraph, LlamaIndex, Pydantic AI, Claude Agents SDK, with knowledge of MCP, context/state management, and agent failure modes.
• Demonstrated evaluation discipline for non‑deterministic systems: building eval sets, defining scoring rubrics (including LLM‑as‑judge approaches), measuring regression with tools like LangSmith, Langfuse, or equivalents, and using evidence to drive architecture decisions.
• Strong backend fundamentals with the depth to drive the tech stack: expert‑level Python, distributed systems, high‑volume data pipelines, and judgment to select and standardize frameworks, libraries, and infrastructure for AI workloads across the team.

Yêu cầu ưu tiên

• Experience in logistics, e‑commerce, or other high‑volume data domains.
• Contributions to open source, technical writing, or conference talks.
• Experience fine‑tuning models (managed fine‑tuning APIs or open‑weight approaches like LoRA/PEFT) and judging when fine‑tuning beats prompting or RAG on cost and quality.
• Experience across the LLM deployment spectrum: working with managed cloud LLM services (AWS Bedrock, Azure OpenAI, Vertex AI, or equivalent) and deploying open‑weight models (e.g., Llama, Qwen from Hugging Face) on your own infrastructure, understanding cost/control trade‑offs.

Mức lương

• Lương: Thỏa thuận

Phúc lợi

• Competitive compensation package
• 13th month salary and ESOP
• Health insurance and annual health check‑up
• Hybrid working, flexible hours, and unlimited leave
• Learning support fund for professional development
• Access to AI tools and innovation initiatives
• Company‑provided laptop & monitor
• Complimentary lunch and daily snacks
• Employee care programs and engagement activities

Hình thức làm việc

• Hybrid working, flexible hours, and unlimited leave

Thông tin khác

• Perform.AI is the AI Commerce Operating System for ambitious brands, building the AI Commerce Operating System for global e‑commerce brands.
• The opportunity to work in a fast‑growing, super exciting, and innovative business building the AI Commerce Operating System for global e‑commerce brands.
• Fully sponsored Claude Code usage, access to AI tools, and a bleeding‑edge agentic development environment.
• Company overview: Founded in 2016 as Parcel Perform, became Perform.AI in September 2026.