Hi, I'm Param
AI Researcher & Machine Learning Engineer
I build agentic AI systems that integrate language, vision, memory and control to create safe, aligned solutions bridging research and real‑world impact.
Param Popat
ML Engineer at Apple
About Me
Passionate about building autonomous systems and advancing AI safety through rigorous research and engineering.
I'm a passionate ML engineer with 4+ years of experience building autonomous systems, photorealistic simulation infrastructure, and on‑device neural network pipelines. My work spans agentic LLM interfaces, NeRF‑based simulators, and RL agents that drive meaningful product impact.
As a machine learning engineer at Apple, I lead the development of Shortcut Agents that transform natural language instructions into executable automations via LoRA‑adapted foundation models and multi‑turn reasoning. Previously, I built city‑scale digital twins with 3D Gaussian Splatting, developed adversarial RL agents, and optimized neural networks for Apple Silicon with pruning and quantization.
I'm committed to advancing AI that is grounded, persistent, and action‑capable. If you're working on frontier LLM architectures or embodied intelligence, I'd love to connect.
Experience
Building the future of AI through innovative research and engineering at leading technology companies.
Machine Learning Engineer (SPG) – Agentic AI & Simulation
Apple Inc., Cupertino, CA
- Led development of Apple's Shortcut Agent, enabling users to describe complex tasks and receive executable automations with >92% success across 100+ templates.
- Built photorealistic simulation infrastructure using 3D Gaussian Splatting; scaled training to 50k+ agent‑hours across 1000+ environments.
- Constructed city‑scale NeRF/3DGS digital twins for closed‑loop training and evaluation of autonomous systems.
- Developed adversarial reinforcement learning agents to simulate long‑tail behaviors, improving robustness of autonomous stacks.
AI Research Intern
AI Zwei (Dailight), New York, NY
- Developed transformer‑based language blocks for production AutoML and applied distillation, pruning, and quantization to boost inference speed and robustness.
Project Trainee – AI
Robert Bosch GmbH, Bengaluru, IN
- Invented and patented methods to detect adversarial attacks and prevent model stealing, commercialized in Bosch's AI security service.
- Built generative recommendation engine that raised conversion rates from 25% to 70%.
Founder
SnapFactory, Ahmedabad, IN
- Founded an AI‑powered photography workflow startup; built a GAN‑based low‑light image enhancer that cut post‑processing time by 60%.
- Managed a team of photographers and freelancers to deliver fast, high‑quality photo delivery across 300+ events.
Machine Learning Intern
Canary Mail, Rajkot, IN
- Developed GRU‑based stock seasonality models, identifying buy/sell windows with simulated 13.5% YoY return.
Education
Academic foundation in computer science with focus on machine learning and artificial intelligence.
M.S. in Computer Science
Columbia University, New York, NY
GPA 3.98/4.0, TA for courses in Computer Vision and NLP.
B.Tech in Computer Engineering
Nirma University, Ahmedabad, IN
GPA 9.12/10.0, TA for Deep Learning.
Projects
Innovative solutions spanning AI research, machine learning infrastructure, and autonomous systems.
Apple Shortcut Agent
A personalized task planner integrated with iOS that combines LoRA‑adapted foundation models with UI context ingestion and multi‑turn feedback loops to generate and refine Apple Shortcuts.
Are Transformers Learning or Memorizing?
Investigated whether language models abstract grammar or simply memorize; trained ALBERT on constrained data and showed equivalence with lookup tables, highlighting limits of current architectures.
Quick‑ML: Email‑Based ML Prototyping
Built a serverless system where users send an email to trigger data ingestion, model training, and endpoint provisioning via AWS Lambda, SES, and SageMaker.
CNN for Next‑Day Stock Prediction
Applied a 2D time‑series CNN to mapped stock indicators, achieving an F1 score of 0.80 in predicting directional movement.
Articles
Thoughts on AI safety, computer vision, and the intersection of technology and society.
Protecting Society from AI‑Generated Misinformation
This Analytics Magazine piece discusses how synthetic content generated by AI can mislead the public and proposes ethical guidelines for detection tools, regulatory oversight and responsible AI deployment.
Read article →Siamese Networks for Visual Tracking
An overview of Siamese neural networks and contrastive loss for real‑time object tracking, demonstrating a fully convolutional Siamese model with a ResNet backbone that achieves 0.88 precision on the OTB dataset.
Read article →The Journey of Mathematics (Layover at Pi)
A personal reflection on why many people dislike formal mathematics and how exploring timeless concepts like π can reveal the discipline's deeper beauty and interconnectedness.
Read article →Publications & Patents
Research contributions in machine learning, computer vision, and AI security.
- Animal/Object Identification Using Deep Learning on Raspberry Pi — published in Springer's Smart Innovation, Systems and Technologies series (2019).
- Method to Protect Neural Networks Against Model Extraction — patent filed in the US, India and Germany (2019).
Get in Touch
If you're interested in collaborating on AI projects, exploring new research directions, or just want to say hello, feel free to reach out.