Scalable ML Systems
Distributed training and inference pipelines that remain reliable and observable at production scale.
Researcher · Engineer
Danish Zulfiqar
Systems & Machine Learning · Electrical & Computer Engineering
I work at the intersection of machine learning and scalable systems, designing intelligent architectures that are both rigorous and deployable. My interests span efficient model inference, distributed pipelines, and applied ML for real-world signals.
About
I am an Electrical & Computer Engineer focused on building intelligent, scalable systems. My work draws on both classical engineering rigor and modern machine learning — from low-level signal processing to large-scale distributed inference.
Outside of research, I contribute to open-source projects, mentor early-career engineers, and occasionally write about systems design. I am always open to thoughtful conversations — feel free to reach out.
News
-
Journal manuscript “Machine learning-based framework for fast calibration of non-linear model predictive control” submitted to Engineering Applications of Artificial Intelligence — currently under review.
-
Presented our poster “Feasibility of an AI-administered COA using a multi-agent framework” at the ISCTM Conference in Amsterdam, The Netherlands.
Research
My research builds principled, efficient systems for learning and inference. I am currently exploring three directions.
Efficient Inference
Quantization, sparsity, and compiler-level techniques that unlock low-latency model serving on commodity hardware.
Applied ML for Signals
Bringing learning methods to physical and temporal signals — sensing, control, and anomaly detection.
Publications
Selected papers and preprints.
-
Journal · Under Review
Machine learning-based framework for fast calibration of non-linear model predictive control
Engineering Applications of Artificial Intelligence · 2026 (Manuscript under review)
@article{zulfiqar2026mlnmpc, title = {Machine learning-based framework for fast calibration of non-linear model predictive control}, author = {Zulfiqar, Danish and Ahmed, A. and Arshad, A. and Ahmed, Q.}, journal = {Engineering Applications of Artificial Intelligence}, year = {2026}, note = {Manuscript under review} } -
Conference · Poster
Feasibility of an AI-administered COA using a multi-agent framework
International Society for CNS Clinical Trials and Methodology (ISCTM) Conference · Amsterdam, The Netherlands · October 10, 2025
@inproceedings{mclaughlin2025coa, title = {Feasibility of an AI-administered COA using a multi-agent framework}, author = {McLaughlin, D. and Zulfiqar, Danish and Haseeb, A. and Hussain, R. and Shakeel, A. and Ahmed, J.}, booktitle = {International Society for CNS Clinical Trials and Methodology (ISCTM) Conference}, address = {Amsterdam, The Netherlands}, note = {Poster presentation}, month = {October}, year = {2025} }
Selected Projects
Engineering work that powers the research — from open-source libraries to applied systems.
Aug 2024 — May 2025
ML-Based Optimal Active Cell Balancing for EV Range Extension
Machine-learning-driven control algorithm that extends EV battery life through optimal active cell balancing. Designed a composite cost function targeting driving-range maximization, and applied three ML techniques — including Transformer-based architectures — to optimize its weights, bridging classical control and learning.
July 2024
Predictive Maintenance Microservice
Microservice for industrial anomaly detection, fault classification, and forecasting of failure metrics such as Fault-to-Active ratio over time. Deployed as a scalable service on AWS using Docker and FastAPI, with SHAP-based model interpretability.
July 2023
Jericho.ai
Retrieval-Augmented Generation (RAG) tool that answers multiple-choice questions across three course materials, designed to enhance learning through fast text search with traceable references back to source content.
Feb 2025
Language Detector (NLP)
NLP model that classifies text across 18 languages. Built a custom dataset scraped from Wikipedia and used a Multinomial Naive Bayes classifier for prediction, with full pipeline from preprocessing to evaluation.
Education
-
2021 — 2025
B.Sc., Electrical & Computer Engineering
Focus on systems, signals, and machine learning.
Contact
The best way to reach me is by email. I read everything and reply to most.
- danishzulfiqar100@gmail.com
- Islamabad, Pakistan
- github.com/danishzulfiqar
- linkedin.com/in/danishzulfiqar