LLM Engineer

Why this role mattersWe’re building production-grade GenAI systems at the core of trading, risk, and enterprise decision-making. This is not a demo lab or a prompt-only role—you’ll design and ship real LLM platforms used by traders, compliance teams, and leadership every day.If you enjoy solving hard problems at the intersection of LLMs, distributed systems, security, and real-world constraints, this role gives you scale, ownership, and impact.What you’ll buildAs a Senior LLM Engineer, you’ll own the end-to-end lifecycle of Large Language Model–powered systems—from architecture to production operations.You’ll work on:- LLM-native platforms for trading intelligence, risk signals, compliance automation, fraud detection, and enterprise knowledge systems- RAG and agentic workflows that integrate market data, internal systems, and secure tools- High-performance, cost-aware inference stacks running in hybrid and multi-cloud environments- Evaluation, safety, and governance frameworks that make LLMs reliable in regulated financial environmentsWhat you’ll doArchitecture & Engineering- Design and ship end-to-end LLM systems: prompt orchestration, RAG pipelines, agents/tool-use, and workflow automation- Build low-latency, scalable APIs and microservices for inference and orchestration- Implement LLMOps/MLOps pipelines: model versioning, prompt CI/CD, experiment tracking, and automated A/B testingData & Retrieval Systems- Engineer secure ingestion pipelines for market data, trades, compliance records, and voice/text transcripts- Design vector search systems with smart chunking, hybrid retrieval (BM25 + embeddings), re-ranking, and auditabilityEvaluation, Safety & Model Risk- Define how “good” looks: hallucination detection, task-level metrics, adversarial testing, and red-teaming- Build human-in-the-loop validation, bias/fairness checks, and explainability where required- Help formalize model risk management for LLMs in productionSecurity & Compliance (Done Right)- Work with Security and Risk teams to embed Zero Trust, least-privilege access, PII controls, and audit trails- Implement guardrails: content filters, policy enforcement, safe tool invocation, and full traceabilityObservability & Cost Engineering- Instrument everything: latency, throughput, token usage, prompt drift, errors, and SLOs- Actively optimize cost using model selection, quantization, caching, batching, and autoscalingTechnical Leadership- Partner with Trading, Risk, Compliance, Infra, and Platform teams- Mentor engineers on LLM patterns, prompt engineering, and production best practices- Contribute to internal platforms, reusable components, and engineering standardsWhat we’re looking forExperience- 5–10+ years in ML / platform / backend engineering- 3+ years building and operating production LLM systemsStrong hands-on skills in:- Languages: Python (primary), plus TypeScript / Go / Java- LLM frameworks: LangChain, LlamaIndex, Semantic Kernel, DSPy (or similar)- RAG systems: FAISS, Milvus, Pinecone; hybrid search; cross-encoder re-ranking- Model ecosystems: OpenAI / Azure OpenAI, Anthropic, Vertex AI; open-source (Llama, Mistral, Phi)- Inference optimization: vLLM, Triton, quantization (GPTQ/AWQ), batching, caching- LLMOps/MLOps: MLflow or W&B, model registries, CI/CD, feature stores- Cloud & infra: Docker, Kubernetes, Terraform, event-driven systems- Observability: Prometheus, Grafana, OpenTelemetry (metrics, logs, traces)- Security fundamentals: OAuth/OIDC, RBAC, secrets management, encryption, data governanceWhy you’ll enjoy working here- Real production impact—no toy demos- Hard engineering problems with clear ownership and scale- Freedom to choose the right models and architectures, not just the trendy ones- A chance to define how GenAI is done responsibly in financial services