Intern - AI Engineering (from Dec 26/Jan 27)

LIGHTHOUSE CANTON PTE. LTD.; Singapore; On-site; SGD 17,000 - 26,000; Full time

Job summary

Lighthouse Canton invites you to join its AI Engineering Internship, operating inside a regulated wealth management environment. You will work in the AI Lab on real internal applications across Investment, Relationship Management, Risk, Operations, and Compliance, under seasoned engineers who guide your growth.

You’ll learn to build production-grade AI systems, from reading code and small fixes to owning clearly scoped features, with structured mentorship and a readout at internship end.

Qualifications

• Programming experience in Python (coursework/projects) plus exposure to JavaScript/TypeScript, Git and the command line.
• Built at least one small project involving LLMs (chatbot, RAG demo, agent, etc.).
• Familiar with agent frameworks and terms like MCP, RAG, evals, guardrails; not required to be expert.

Responsibilities

• Read and modify production-grade AI code; start with small fixes and progress to owning features.
• Assist in building and tuning retrieval systems: embeddings, vector search, re-ranking.
• Practice prompt engineering and context management to maintain agent coherence across steps.
• Wire internal knowledge bases, market data feeds, and ops systems into agents via MCP servers and typed-tool interfaces.

Skills

Python

JavaScript/TypeScript

Git

Command line

LLMs

RAG

Agent frameworks

LangChain

LangGraph

CrewAI

OpenAI Agents SDK

Claude Agent SDK

MCP

RAG evals

Guardrails

Codex / Copilot

SQL/NoSQL

FastAPI/Flask

Education

Bachelor's/Master's in CS/AI/DS/SE/Quant Finance

No prior professional experience

Strong coursework/open-source/Projects

Penultimate-year or final-year

Tools

None specified

Job description

About Lighthouse Canton

Founded in 2014, Lighthouse Canton is a Singapore-headquartered global investment management firm with approximately US$ 6+ billion in Assets Under Management (AUM). Lighthouse Canton is one of the leading and award-winning investment firms in the region, focusing on managing funds with uncorrelated investment strategies, providing investors with the opportunity to achieve consistent returns across market cycles. Lighthouse Canton has offices in Singapore, Dubai, India, and the UK.

Job Purpose

We are offering an AI Engineering Internship designed as a structured, mentored introduction to building production-grade AI systems inside a regulated wealth management firm. You will sit inside our small AI Lab and work alongside experienced engineers on real internal applications across Investment, Relationship Management, Risk, Operations, and Compliance - not toy projects, and not coffee runs.

What You Will Do

Building & Supporting AI Applications

• Pair with senior engineers on AI-powered internal applications using LLMs, RAG pipelines, and agents. Expect to start by reading code and small fixes, and ramp up to owning well-scoped features under guidance.
• Help build and tune retrieval systems: chunking strategies, embeddings, vector search, and basic re-ranking.
• Practise prompt engineering and context engineering - structuring inputs, managing memory, and compacting context - to keep agents coherent across multi-step tasks.
• Help wire internal knowledge bases, market data feeds, and operational systems into agents via Model Context Protocol (MCP) servers and other typed-tool interfaces.

Agent Harness & Evaluation

• Get hands-on with the agent harness - the runtime layer around the LLM that manages execution loops, tool calls, state, memory, and error handling. You will be expected to learn the vocabulary (ReAct, planner/executor, guardrails, permission gates, loop budgets) and apply it in real code.
• Help build evaluation harnesses: writing test cases, regression suites, deterministic checks (schema validation, citation checks),and simple LLM-as-judge graders - so we can tell whether changes actually improve quality.
• Instrument agents with basic observability: tracing tool calls, logging token usage, latency, and outcomes; help triage failure traces and feed lessons back into prompts and tool specs.
• Contribute to internal documentation on harness patterns, evals, and lessons learned - useful both to the team and to your own learning.

Collaboration & Knowledge Sharing

• Partner with stakeholders in Investment, Relationship Management, Risk, Operations, and Compliance to understand requirements - always alongside a senior team member.
• Take part in code reviews, design discussions, and weekly team check-ins. Asking questions is part of the job, not a sign of weakness.
• Present a short end-of-internship readout summarising what you built, what you learned, and what you would do differently.

What We Are Looking For

Technical Foundations

• Programming: Comfortable in Python from coursework, personal projects, or prior internships. Any exposure to JavaScript/TypeScript, Git, and the command line is a plus.
• AI/ML Curiosity: Has built at least one small project involving LLMs (e.g., via OpenAI, Anthropic, or open-source models) - a chatbot, a RAG demo, a hackathon project, a tinkering side project, an agent that does something silly. Quality of curiosity matters more than scale.
• Agentic AI Awareness: Familiar at a conceptual level with agent frameworks (e.g., LangChain, LangGraph, CrewAI, OpenAI Agents SDK, Claude Agent SDK) and modern terms like MCP, RAG, evals, and guardrails. We do not expect mastery - we expect that you have read about them and want to go deeper.
• AI-Assisted Development: Comfortable using coding assistants such as Claude Code, Codex, or GitHub Copilot as part of how you build - and aware that they need supervision, tests, and clear specs to produce reliable code.
• Data & Web Basics: Light exposure to databases (SQL, or any NoSQL store), and to web/API basics (HTTP, JSON, a back-end framework such as FastAPI or Flask) is helpful but not required.

Mindset & Soft Skills

• Genuine curiosity about AI, finance, and the messy intersection of the two. You should be excited to learn how a wealth management firm actually works.
• Comfort with ambiguity: real projects rarely arrive with neat specs. We will help you scope them, but we want someone willing to sit with uncertainty.
• Communication: able to explain technical ideas in plain English (and in writing), and to flag blockers early rather than getting stuck quietly.
• Integrity: this is a regulated environment. We need an intern who treats client data, MNPI, and internal information with the same care a full-time hire would.

Education

• Currently pursuing a Bachelor's or Master's degree in Computer Science, Artificial Intelligence, Data Science, Software Engineering, Quantitative Finance, or a related field—or a recent graduate within 12 months.
• No prior professional experience required. Strong coursework, personal projects, open-source contributions, hackathon work, or research projects are all valid signals.
• Penultimate-year and final-year students are encouraged to apply.