AI engineer in Sydney, Australia

Jaideep Singh Garlyal

Building production agentic AI at Purple TIC Enterprises.

I build AI systems that turn messy business workflows into useful products.

I’m an AI engineer and data science graduate student in Sydney, working across agentic AI, RAG systems, automation, backend engineering, and applied finance. Before models, I learned sales. Before dashboards, I learned accountability. I care about systems, incentives, and people.

01

  • Agentic AI systems

    Multi-agent workflows, RAG pipelines, and internal automation tools for real business operations.

  • AI × Finance

    Research, modelling, and analytical systems shaped by equity research and financial analysis.

  • Sales × Systems

    A practical operating lens built from closing high-value sales and working directly with customers.

02

  • QuantOnion

    An AI agent for quant research. Ask it about any ticker and it works out the current market regime, compares strategies, runs ML price forecasts, and pulls the risk numbers.

    PythonAgentic AIConnectOnionHMMscikit-learnyfinance

  • Astrion DQ: Agentic Data Quality Triage

    Synogize · UTS Innovation Lab · 2026

    An agentic system that detects, ranks, and explains data-warehouse problems (duplicate transactions, broken foreign keys, invalid dates, nulls, and other silent issues), ordering them by business impact instead of just technical severity. Tested on a retail star-schema warehouse with injected issues and a documented ground truth, it hit an F1 of 0.857 with zero false positives and a 9-second runtime.

    PythonLangGraphDuckDBSQLpandasStreamlitpytest

  • Legal Statute Identification (LeSICiN)

    Pulling the right legal statutes out of case text by combining what the text says with how cases and statutes link together. It's the research that fed into my paper on conversational AI for legal summarisation.

    PythonNLPGraph MLDockerJupyter

  • Anomaly Detection for Audit

    Finding the purchase-order transactions an auditor should actually look at, using Benford's Law, PCA, and unsupervised clustering to flag the odd ones out.

    Pythonscikit-learnPCAClusteringPandas

  • Applied NLP & Data Analysis

    A set of notebooks where I worked through document parsing, embeddings, clustering, and recommendations on messy real-world text.

    PythonTransformersspaCyKerasPandas

03

  1. 01
    Purple TIC EnterprisesAI Engineer, Founding Team
  2. 02
    SynogizeAgentic AI Apprentice, UTS Innovation Lab
  3. 03
    KOSECEquity Research Analyst
  4. 04
    Harbour LaneLead Sales Analyst
  5. 05
    Avkalan.aiAI Engineer
  6. 06
    National University of SingaporeResearch Scholar
  7. 07
    University of Technology SydneyMaster of Data Science and Innovation

PubAn Analysis on Integrating Advanced Conversational AI in Legal Summarization and Information Retrieval ORCID ↗

04

If something here clicks, open a channel.

I'm on Sydney time, usually thinking about systems, incentives, and people. I read everything that isn't a robot.

Freq