Jay Malhotra
Jay MalhotraJay MalhotraAI/ML engineer, MMA at McGill
Open to internships and full-time roles in Toronto or remote in Canada

I build AI products end to end,from the model to the app people use.

  1. Model

    • LangGraph
    • LangChain
    • RAG
    • scikit-learn
    • YOLO, U-Net

    Seen inAsOfSkillGuardKidneyStone-X

  2. Data

    • IBM Db2
    • PostgreSQL
    • SQL Server
    • pandas
    • Vector search

    Seen inOlist late deliveries

  3. API

    • FastAPI
    • Django
    • Flask
    • Node.js
    • AWS Bedrock

    Seen inPerfect Pair

  4. App

    • Flutter
    • Next.js
    • Admin dashboards
    • Encrypted chat

    Seen inAyebot

Written in Python, SQL, JavaScript and C#. Shipped with Docker, Git, AWS and Google Cloud.

Selected work

Six things I built or led. Each one opens a short case study with real footage, the numbers and the limits.

Nov 2026IBM Db2Hack 2026, 2nd place, solo

AsOf

Checks an AI assistant's answers against the policy in force on the customer's date.

0 of 65 wrong drafts released on blind setsRead the case study
Oct 2026BSRM x ALDO Retail Gen AI Hackathon, team of five

Perfect Pair

An agentic returns assistant for a footwear retailer.

prize winnerRead the case study
Oct 2026Nebius x NVIDIA Global AI Hackathon, team of two

SkillGuard

Tests what a new skill does before an AI agent can install it.

blocked a skill a static scanner rated safeRead the case study
Fall 2026McGill INSY 674, led a team of seven

Olist late deliveries

Flags e-commerce orders likely to arrive late, at the moment they are approved.

top 10% of flags catch 36% of late ordersRead the case study
Nov 2025Bennett University course project, team of three

KidneyStone-X

Finds kidney stones on CT scans and outlines the kidney around them.

mAP@0.5 of 0.965 on validationRead the case study
Since Jan 2026Founding engineer

Ayebot

An AI assistant and encrypted messaging platform for universities.

in production, pilot signedRead the case study

Try one

The idea behind AsOf, live. An AI assistant gives two customers the same answer. Whether it is right depends on which policy version applied on each customer's date.

Customer's date
10 March 2026
AI draft, the same on every date

“Yes, you can still apply for a bereavement fare after your trip.”

Policy in force on that date: BRV-002 v1

Drag the date and watch the verdict flip. Read the AsOf case study.

Experience

AyebotJan 2026 to presentFounding engineer
Trench GroupSep 2026 to presentData engineer, McGill MMA capstone project
PerpexAug to Nov 2025Software development intern
SmartInternzJul to Aug 2025Google Cloud generative AI intern

Education

McGill University, DesautelsIn progressMaster of Management in Analytics
Bennett University2023 to 2026BCA with AI specialization, CGPA 9.28/10

More work

LLM Evaluation PipelineScores chatbot responses for relevance, hallucination and completeness, with latency and cost tracked per stage.
Risk-factor disclosure researchTests whether year-over-year changes in the Item 1A risk-factor text of SEC filings forecast volatility and crash risk. In progress, with a co-author from finance.

From the model to the app.

I like owning a problem from the model to the interface a person actually uses. In the last year that has meant a production platform for universities, three hackathon builds this fall, and a machine learning project I led with a team of seven.

I grew up in India and did a BCA with an AI specialization at Bennett University. I am now in Canada for the Master of Management in Analytics at McGill, which adds the business and data side to the engineering.

I am looking for an AI/ML engineering, data engineering or software role, as an internship or full time, in Toronto or remote in Canada.

Have a role in mind? Let's talk.

jaymalhotra106@gmail.com