Get a reply from this recruiter
Your resume and cover letter, tailored to exactly what they’re hiring for.
Lancesoft is seeking an experienced analytics/AI engineer to design and implement multi-agent workflows, integrating tools and APIs via standardized contracts. You will work with LangGraph, LangChain, and various SDKs to orchestrate tool use and ensure robust, deterministic actions in complex business processes.
You will contribute to RAG designs, context handling, and the evaluation of model behavior across different weights and providers.
Required SkillsStrong understanding of agent loops (ReAct, plan-and-execute) and how models decide between calling a tool and answering directly, with hands-on exposure to frameworks such as LangGraph, LangChain, OpenAI Agents SDK, Claude Agent SDK, or PydanticAI Working knowledge of multi-agent patterns (orchestrator/worker, router, supervisor) and handoffs, using LangGraph, CrewAI, or AutoGen Experience defining function-calling specs and typed tool contracts using JSON Schema, Pydantic, and OpenAPI to turn user intents into deterministic actions Knowledge of Model Context Protocol (MCP) for exposing tools and data to agents, and awareness of A2A for agent-to-agent communication Experience modeling multi-step workflows as state machines or graphs with checkpoints and resume logic, using LangGraph state management or durable workflow engines such as TemporalAbility to specify idempotent actions, retries, timeouts, and fallback handling for validation failures, partial results, and ambiguous intents Understanding of LLM fundamentals across Claude, GPT, Gemini, and open-weight models (Llama, Mistral, Qwen): context windows, structured outputs, sampling, and failure modes such as hallucination and tool misuseWorking knowledge of RAG design, including embeddings, hybrid search, chunking, and re-ranking, using vector stores such as pgvector, Pinecone, Qdrant, or OpenSearch Understanding of context engineering and cost/latency levers: token budgets, memory and compaction, prompt caching, and model routing Experience building agent eval datasets and grading criteria for tool selection, parameter extraction, task completion, and hallucination rates, using LangSmith, Langfuse, Arize Phoenix, Braintrust, or Promptfoo Ability to define tracing, telemetry, and audit lineage requirements using OpenTelemetry, LangSmith, or Datadog Experience designing human-in-the-loop approval tiers and guardrails (NeMo Guardrails, Guardrails AI, or custom policy layers), plus awareness of prompt injection, data leakage, and least-privilege access via OAuth 2.0 and scoped credentials Ability to write Python and SQL to prototype tool calls, mock responses, and inspect LLM outputs5+ years as a business analyst gathering and documenting requirements directly with business users, product, and engineering 2-3 years of hands-on experience designing or delivering agentic AI/LLM-based solutions, including agent workflows, tool integration, human-in-the-loop controls, and AI guardrails.Strong skills in writing user stories, acceptance criteria, process flows, and functional and API specifications Experience working in agile teams, managing backlogs in Jira and Confluence, and prioritizing features with product owners Ability to translate business goals into measurable success metrics and explain technical trade-offs to non-technical stakeholders
Knowledge of buy-side trading workflows: order lifecycle, order staging, routing, execution, allocations, and post-trade Understanding of portfolio management workflows: portfolio construction, rebalancing, model portfolios, cash management, and what-if analysisFamiliarity with pre-trade and post-trade compliance, restricted lists, and best execution requirements Experience with OMS/EMS platforms such as Charles River, Aladdin.Exposure to multiple asset classes, such as equities, fixed income, FX, or derivativesBachelor's degree in Business, Finance, Computer Science, or related field