Hello, I'm

Muhammad Hashir Bin Khalid

AI/ML Engineer & Product Architect

I design and ship AI systems end-to-end — computer vision, LLMs & agentic architectures, and the data pipelines behind them.

View My Work Download CV
Muhammad Hashir Bin Khalid

About Me

I'm a Pakistani Fulbrighter with a Master's in Computer Science from New York University — an AI/ML engineer and product architect with ~9 years building intelligent systems. I've contributed to commercial AI products across computer vision, NLP, and IoT, and I currently work on agentic AI harnesses, LLM systems, and large-scale BigQuery data pipelines. I'm passionate about using technology to improve human lifestyle.

  • Name Muhammad Hashir Bin Khalid
  • e-mail mohammadhashirbinkhalid@gmail.com
Download Resume Contact Me

What I Do

I design and ship AI systems end-to-end — and increasingly architect agentic AI harnesses and advise teams on building real-world AI products.

Agentic AI & LLM Systems

Designing agentic harnesses, multi-agent orchestration, and LLM-powered systems — from intent understanding to tool-using, autonomous workflows.

AI/ML & Deep Learning

End-to-end machine learning and deep learning — data, training, evaluation, and deploying models that hold up in production.

Computer Vision

Detection, segmentation, classification, medical imaging, and emotion recognition.

Data Pipelines & BigQuery

Large-scale data pipelines, warehousing, and analytics on BigQuery — turning raw data into queryable intelligence.

Embedded Systems & IoT

Firmware, control circuits, sensing, and connected devices that bridge AI and hardware.

AI Systems Consulting & Architecture

Architecting AI products and advising teams from research prototype to shipped, real-world system.

Skills & Tech Stack

AI / ML

TensorFlowKerasPyTorchscikit-learnDeep LearningClassical ML

Domains

Computer VisionNLP & LLMsAgentic AIForecasting & OptimizationMedical Imaging

Data

BigQueryHive / HadoopData Pipelines

Embedded / IoT

Embedded CMCU / FirmwareSensorsWi-Fi / BLE

Languages

PythonCC++JavaMATLAB

  • Education

  • MSc in Computer Science - New York University

    2020 - 2022

    Completed my 2-year Master's degree in Computer Science on the Fulbright Scholarship

  • B.E. in Electronic Engineering - NEDUET

    2014 - 2017

    Completed my 4-year Bachelor's degree in Electronic Engineering with Distinction (Gold Medalist)

  • Experience

  • Senior ML Engineer / Product Architect - Hyly.AI

    Dec 2024 - Present

    Architecting AI products and agentic systems — designing agentic harnesses, LLM-powered workflows, intent understanding, and the BigQuery data pipelines behind them.

  • ML Engineer - Hyly.AI

    Dec 2023 - Nov 2024

    Developed AI/ML and NLP features end-to-end — model training, LLM fine-tuning, and turning data into product capabilities.

  • Team Lead - NCL NEDUET.

    Aug 2022 - Dec 2023

    Technical lead across an electric-mobility ecosystem for AIM Motors / AIM-GE LLC (USA) — directing AI, firmware, and software teams spanning the e-bike, swappable batteries with onboard BMS, and a battery-swapping network.

  • Senior Machine Learning Engineer - Lambda Theta

    Sep 2022 - Nov 2023

    Led AI and computer-vision work on local and international projects — including road-condition monitoring with ELM (crack detection, classification, and severity estimation using deep learning and computer vision).

  • Research Associate - NCL National Center of AI

    Sep 2020 - Aug 2021

    Led a team across machine learning, AI, firmware, and embedded systems — including a chest X-ray lung-trauma predictor with Aga Khan University Hospital and intelligent ceiling-fan R&D.

  • Artificial Intelligence Engineer - Instrumentation Center, NED

    Dec 2017 - Aug 2020

    Built the IoT control circuit and trained the AI system for the HEC-funded Smart Fans project — embedded sensing, ML usage modeling, and energy-saving control.

Awards & Honors

Fulbright Scholarship

Fully-funded U.S. scholarship for the Master's at New York University.

NYU GSAS Full Tuition Scholarship

Graduate School of Arts & Science full-tuition award for the Master's.

Gold Medal — NED University

Highest CGPA in the graduating batch.

Gold Medal — Institution of Engineers Pakistan

1st position in undergraduate studies.

Huawei Certified — Artificial Intelligence

Certified ICT Associate in AI, and AI Academy Instructor (Presidential Nominee).

Studied under Yann LeCun

Deep Learning at New York University.

Publications

7 peer-reviewed publications · IEEE · Elsevier · MDPI · 2021–2026

Scalable Optimization of Ultra-Dense Heterogeneous Networks Using Stochastic Geometry and Deep Learning Techniques

Amna Shabbir, Muhammad Hashir Bin Khalid, Hashim Raza Khan, Kamran Arshad, Khaled Assaleh

Combines stochastic geometry with deep learning to optimize energy efficiency in ultra-dense 5G networks under practical QoS constraints.

AI, vol. 7, no. 2, art. 76 2026 · MDPI

Deep LearningWirelessEnergy Efficiency
Read paper →

Fuzzy MRAS Speed Sensorless Induction Motor Drive Control for Electric Vehicles

Saqib Jamshed Rind, Saba Javed, Hashim Raza Khan, Muhammad Hashir Bin Khalid, Kamran Arshad, Khaled Assaleh

A fuzzy-logic MRAS observer for sensorless induction-motor speed estimation in EVs — robust to load-torque disturbances.

Energies, vol. 19, no. 6, art. 1580 2026 · MDPI

EVMotor ControlFuzzy Logic
Read paper →

Design and Development of a Configurable High Voltage Power Supply for Atmospheric Plasma Generation

Taha Amjad, Muhammad Hashir Bin Khalid, Mubbshir Saleem, Hashim Raza Khan, Kamran Arshad, Khaled Assaleh

STM32-controlled push-pull inverter delivering an adjustable output up to ~20 kV for atmospheric plasma, with over-voltage and over-current protection.

2026 1st International Conference on Emerging Technologies and Engineering Systems (ICETES), pp. 348–353 2026 · IEEE

Power ElectronicsEmbeddedHigh Voltage
Read paper →

A Low-Cost Energy Monitoring System with Universal Compatibility and Real-Time Visualization for Enhanced Accessibility and Power Savings

Hashim Raza Khan, Majida Kazmi, Lubaba, Muhammad Hashir Bin Khalid, Urooj Alam, Kamran Arshad, Khaled Assaleh, Saad Ahmed Qazi

A low-cost, universally-compatible IoT energy monitor with a secure real-time dashboard — about 40% cheaper to build than comparable systems.

Sustainability, vol. 16, no. 10, art. 4137 2024 · MDPI

IoTEnergy MonitoringEmbedded
Read paper →

Dataset of Usage Pattern and Energy Analysis of an Internet of Things-Enabled Ceiling Fan

Hashim Raza Khan, Muhammad Hashir Bin Khalid, Urooj Alam, Mahnoor Atiq, Uvais Qidwai, Saad Ahmed Qazi

An open dataset logging BLDC ceiling-fan usage with temperature and humidity over Wi-Fi to enable energy-pattern analysis.

Data in Brief, vol. 46, art. 108900 2023 · Elsevier

IoTDatasetEnergy
Read paper →

Threefold Optimized Forecasting of Electricity Consumption in Higher Education Institutions

Majida Kazmi, Hashim Raza Khan, Muhammad Hashir Bin Khalid, Saad Ahmed Qazi

Compares ANN vs Random Forest for building load forecasting — Random Forest reaches 2.42% MAPE in real time.

Computers, Materials & Continua, vol. 73, no. 2 2022

Machine LearningForecastingEnergy
Read paper →

An Isolated Power Factor Corrected Cuk Converter with Integrated Magnetics for Brushless DC Ceiling Fan Applications

Hashim Raza Khan, Majida Kazmi, Haris Bin Ashraf, Muhammad Hashir Bin Khalid, Abul Hasan, Saad Ahmed Qazi

An isolated PFC Cuk converter with integrated magnetics for sensorless BLDC fan drives — cuts motor torque ripple to 2.14% at 0.999 power factor.

Electronics, vol. 10, no. 14, art. 1720 2021 · MDPI

Power ElectronicsMotor DrivesPFC
Read paper →
View on Google Scholar

Projects

Products & research · Computer Vision · NLP & Agentic AI · IoT · EV

Deployed

RoboChotu — Autism Assistant Robot

Contributed to RoboChotu, a social robot that helps therapists teach autistic children — emotion recognition and a multilingual chatbot. Tested at 3–5 autism centers; HEC/NCAI-funded.

AIRoboticsHealthcare
Read more
  • Co-founded as the founding student team's project head; later product architect.
  • Worked on the on-device emotion-recognition model (CNNs + classical ML) and the robot firmware (Arduino locomotion + Android-tablet face/OS, Bluetooth).
  • Built activity workflows, the multilingual chatbot (English/Urdu/Arabic), and the in-class experiment design.
  • Won multiple HEC/NCAI research & project grants; tested with ~10–50 children across 3–5 autism centers.
Page
Deployed

SmartFan — AI Ceiling Fan

Contributed to SmartFan, an AI ceiling fan that learns usage and saves energy — including the ML models and the MCU control circuit. Deployed across residential and commercial sites, saving thousands of kWh.

IoTAIEnergy
Read more
  • Worked on the ML models (preferred speed, occupancy on/off, energy optimization, usage scheduling) using classical ML, neural networks, and time-series.
  • Designed the MCU control circuit and sensor integration; built the IoT data-collection pipeline.
  • Co-authored the published usage dataset (Data in Brief, 2023).
  • Grew from the HEC-funded "Intelligent Ceiling Fans" project; deployed across residential & commercial sites.
Product
Deployed

INSTRUX — Energy Monitoring

Contributed to INSTRUX, a real-time energy-monitoring platform — including the embedded metering hardware. Deployed at multiple sites including NED University, Oh My Grill, and Rainbow Textile.

IoTEmbeddedEnergy
Read more
  • Worked on the embedded metering hardware — analyzer-interface board, MCU firmware, sensing & calibration, and ~20-second Wi-Fi telemetry — plus usage trend analysis.
  • Hands-on with the system behind the Sustainability (2024) paper; ~40% cheaper to build than comparable systems.
  • Deployed at NED University, Oh My Grill, and Rainbow Textile.
Product Paper

Road Condition Monitoring & Crack Detection

Contributed to a deep-learning road-condition system — crack detection, classification, and severity estimation on large real-world road imagery (delivered to ELM).

Computer VisionDeep Learning
Read more
  • Built core computer-vision models: crack detection, crack-type classification, and severity estimation.
  • Techniques: CNN segmentation, object detection, classification networks, and classical CV.
  • Trained and tested on a large real-world road-imagery dataset; delivered to ELM.

AIM Motors EV Ecosystem

Contributed to a full electric-mobility solution — e-bike, swappable battery with BMS, anti-theft pairing, and a city swapping network, with AI for range prediction and battery-supply planning. Delivered to AIM-GE (USA) and field-tested.

EVEmbeddedAI
Read more

A full electric-mobility solution — not a single product — spanning:

  • Electric bike + onboard firmware / control circuit.
  • Swappable battery + firmware / control circuit with onboard BMS.
  • Anti-theft pairing — the bike pairs with its battery's circuit.
  • City-wide swapping stations for battery exchange.
  • AI — range prediction (remaining km) and battery-supply planning across stations (fleet / logistics optimization).

Role: Technical Team Lead across AI, firmware, software, and electronics. Outcome: a small prototype fleet (2–10 bikes) + a proof-of-concept swap station; delivered to AIM-GE LLC (USA) and field-tested.

Robothon — Meta-ML Platform

Contributed to Robothon, an NYU capstone meta-ML platform where ML bots compete to predict ADHD/ASD stress and recommend an Apple Watch activity.

Machine LearningAutoMLHealthcare
Read more
  • NYU master's capstone; sole builder.
  • Built a meta-ML competition framework where ML "bots" compete to produce the best model.
  • Tech: classical ML + an AutoML / meta-learning design.
  • Applied to predicting ADHD/ASD stress and recommending an Apple Watch activity.

Machine Common Sense (DARPA)

Contributed to NYU's part of DARPA's Machine Common Sense program — computer-vision and simulation work (Prof. Rob Fergus).

AIComputer VisionResearch
Read more
  • Master's capstone contributing to NYU's part of DARPA's Machine Common Sense program (Prof. Rob Fergus's group).
  • Worked on simulations and computer-vision models for scene evaluation, supporting the larger research team.
DARPA program

Lung Trauma Injury Predictor

Contributed to a chest-X-ray lung-trauma injury predictor with Aga Khan University Hospital — deep learning with transfer learning, with promising accuracy.

Computer VisionMedical ImagingHealthcare
Read more
  • Led the AI team building a chest X-ray lung-trauma injury predictor with Aga Khan University Hospital.
  • Tech: CNNs + transfer learning + medical-image preprocessing on a medium dataset (1k–10k X-rays).
  • Promising accuracy; later continued by the team.
7

Publications

3

Products Contributed To

8

Projects Delivered

9

Years in AI / ML

Get in Touch

Interested in working together, or have a question? I'd love to hear from you.

mohammadhashirbinkhalid@gmail.com

Email Me