Junior AI Engineer

eClerx•New York, NY

3d•$100,000 - $120,000•Onsite

About The Position

eClerx supports global financial institutions across trade support, financial crime compliance, client lifecycle, asset servicing, settlements & clearing, and data/document operations—often at large scale and under strict accuracy and compliance expectations. You’ll join a team building AI-powered tools (agentic workflows + GenAI capabilities) that help streamline operations, improve control effectiveness, and reduce manual effort—while remaining audit-ready and secure.

Requirements

2+ years’ experience building production software with strong hands-on coding ownership.
Strong Python programming skills with comfort working in a shared professional codebase; familiarity with Java or TypeScript/JavaScript is a plus
Foundational understanding of LLMs: attention mechanisms, tokenization, context windows, and inference trade-offs
Practical experience with LLM/GenAI development: prompt and context design, structured outputs, tool calling, and agentic workflows
Familiarity with common engineering practices: Git, REST APIs, logging basics, and data handling (JSON, SQL)
Ability to work in ambiguous problem spaces and iterate quickly with feedback from product and ops stakeholders

Nice To Haves

Personal projects, open-source contributions, or research on ML or applied ML
Experience building RAG systems: chunking strategies, embedding models, vector stores (e.g. FAISS), and retrieval evaluation
Hands-on experience with agentic orchestration frameworks such as LangGraph or LangChain-style state machines, deployed into real APIs or services
Exposure to financial services operations (capital markets ops, KYC/ client lifecycle, AML/compliance, reconciliations, exceptions processing).
Experience with intelligent document processing pipelines: PDF parsing, schema mapping, and field validation
Familiarity with deep learning frameworks (PyTorch, TensorFlow, or JAX) and strong applied ML fundamentals
Knowledge of responsible AI concepts: privacy, explainability, and evaluation discipline
Full-stack development experience (e.g. React, Next.js, FastAPI) is a meaningful bonus.

Responsibilities

Build agentic & GenAI tools for Financial Markets use cases
Develop agentic AI workflows that can plan, execute, and validate steps across operational processes (e.g., exception triage, document checks, workflow routing, and knowledge-assisted decisioning).
Build GenAI features such as summarization, classification, extraction, drafting, and Q&A for financial operations, with a focus on reliability and traceability.
Create RAG pipelines to ground LLM outputs in approved knowledge sources (policies, SOPs, regulatory artifacts, client documentation, historical cases).
Engineer production-grade integrations
Implement tool-calling / function-calling patterns so agents can safely interact with internal systems (case management, CRMs/ERPs, data services, document repositories) and produce structured outputs.
Build backend services and APIs that package AI capabilities into reusable components (internal tools and/or client-facing solutions).
Write clean, maintainable code with strong engineering hygiene: unit tests, integration tests, code review participation, CI/CD readiness, and documentation.
Quality, evaluation, and responsible deployment
Design lightweight evaluation harnesses and test suites (golden datasets, regression tests, prompt/version tracking, and error analysis) to continuously improve accuracy and consistency.
Implement guardrails and human-in-the-loop checkpoints where appropriate (especially for compliance-sensitive decisions).
Collaborate with domain SMEs (ops, compliance, risk) to translate requirements into measurable system behavior and acceptance criteria.