Rengo AI is building the intelligence layer for fund management — starting with next-generation portfolio monitoring systems for investment teams.

Today, portfolio monitoring is fragmented across dashboards, spreadsheets, internal tools, and manual analyst workflows. Rengo replaces this with an AI-native monitoring layer that continuously interprets portfolio activity, risk, exposure, and performance across assets and strategies.

The Role

As a Founding AI Engineer, you will build the core system that powers AI-driven portfolio monitoring for institutional investors.

You will design systems that continuously:

ingest portfolio + market + position-level data

detect meaningful changes and anomalies

generate structured investment insights

explain performance and risk drivers in natural language + structured outputs

This is a high-reliability AI system, not a chatbot.

What You’ll Build

1. AI Portfolio Monitoring Engine

Real-time and batch systems that monitor:

portfolio performance (PnL, attribution, drawdowns)

exposure shifts (sector, geography, asset class)

risk signals (volatility, correlation, concentration)

position-level changes

AI layer that converts raw portfolio data into:

alerts

summaries

explanations

actionable insights

2. Change Detection & Intelligence Layer

Build systems that detect:

significant portfolio movements

abnormal price/volume behavior in holdings

drift from target allocations

risk regime changes

Prioritization layer: what matters vs noise

3. AI-Generated Portfolio Narratives

Generate structured outputs such as:

daily / weekly portfolio reports

performance explanations (“why did we lose/gain?”)

exposure breakdowns

risk commentary

Ensure outputs are:

auditable

grounded in data

consistent across runs

4. Data + Retrieval Systems for Funds

Integrate:

positions & holdings data

market data feeds

internal fund metadata

external news & filings (optional enrichment layer)

Build RAG pipelines over portfolio + market context

5. LLM Systems for Financial Reliability

Design LLM pipelines that:

avoid hallucinated financial reasoning

produce structured, verifiable outputs

ground insights in actual portfolio data

Build evaluation frameworks for correctness of financial narratives

Requirements

Strong engineering background

3–7+ years in backend, data engineering, or ML systems

Strong Python (mandatory)

Experience building production data systems or analytics platforms

LLM / AI systems experience

Experience building LLM applications in production

Strong understanding of:

RAG systems

structured generation (schemas, JSON outputs)

tool use / function calling

agent workflows

Awareness of failure modes in LLM reasoning (critical in finance)

Data-heavy systems mindset

Experience with:

time-series data

event-driven pipelines

analytics / observability systems

Comfort working with imperfect, high-volume financial data

Nice to Have

Experience in:

asset management / hedge funds / fintech

portfolio analytics or risk systems

trading / market data infrastructure

Familiarity with:

exposure/risk models

PnL attribution systems

BI / analytics platforms for finance

Experience with vector databases or hybrid retrieval systems

What Makes This Role Unique

You are building the core monitoring brain of a fund

Not dashboards — interpretation + intelligence

Systems you build directly influence investment decisions and risk awareness

High emphasis on:

correctness

traceability

reliability under uncertainty

You own the full stack: data → intelligence → insight delivery

Tech Direction

Python (core systems + AI orchestration)

LLM APIs (OpenAI / Anthropic / open-source models)

Postgres + time-series storage

Vector DB for semantic retrieval

Stream/batch processing pipelines

Cloud infrastructure (AWS/GCP)

Why Join

Define how AI monitors institutional portfolios

Replace manual analyst workflows with automated intelligence systems

Work on one of the hardest AI problems in finance: turning data into trustworthy interpretation

High ownership, early-stage, no legacy constraints