Ishu Garg
I build AI systems that turn complex research into repeatable workflows. My work spans agentic research platforms, enterprise copilots, and deep learning for financial data.
Experience at Bezant, Millennium, Microsoft, and Goldman Sachs.
Research, applied AI, and the systems that connect them.
Bezant
Tech Lead
New York, US · Remote from India
Concept to production
Took an agentic research and backtesting platform from concept to production. Designed and built reproducible research workflows, structured agent interfaces, and validation checks for experiments.
Microsoft
Data & Applied Scientist II (Applied AI)
Led work on Sales Copilot, building LLM-based workflows for lead discovery and pitch generation. Also developed machine-learning systems for partner recommendations.
Millennium
Quantitative Researcher
New York, US · Remote from India
Developed proprietary deep-learning methods for financial problems with limited labeled data, alongside quantitative research using fundamental, price-volume, and alternative data.
A foundation in computer science and machine learning.
IIT Madras · 2013–2018
B.Tech + M.Tech in Computer Science and Engineering. Branch change to Computer Science. Ranked first in the freshman cohort. Star Teaching Assistant Award.
Carnegie Mellon · 2017
Research intern, Machine Learning Department. Geometry-aware deep learning for unsupervised depth and motion estimation from monocular video. S.N. Bose Scholar.
Technical judgment, a few good questions, and space for poetry.
Featured poem
I built myself in layers, slowly, like a wall learning its own weight.
All poemsTool mistakes, source boundaries, and small tests that make failures inspectable.
Read Evals Before AgentsAlgorithms, proofs, and the pleasure of seeing an idea click.
Looking for something practical? Learning resources and puzzles have their own corners.
For AI systems, research engineering, and agentic workflow work.