00 / RESEARCH RECORD WEB CV

RESEARCH & PROFESSIONAL RECORD

Resume PDF ↓ Scholar ↗

Current research, publications, intellectual property, funded work, appointments, teaching and academic service.

01/CURRENT_RESEARCH
01 ACTIVE / 01 UPCOMING
01REWIREActive
Grid World Models / Critical Infrastructure
SinceAug 2026 DomainElectricity Grid

Topology-aware world model for electricity-grid reconfiguration that evaluates line-switching candidates through learned imagined rollouts and verifies promising actions using exact AC power-flow simulation — covering topology-changing actions, ranking of candidates, and identification of unsafe or poor candidates. Experiments are ongoing; no performance claims are made.

02INTACTStarting Sep 2026
LLM Reliability
StartSep 2026 SignalGeneration-Time

Investigating whether generation-time signals from autoregressive language models — particularly token-level logit and probability structure — can provide reliable evidence of output correctness, whether these signals distinguish reliable from incorrect or unsupported outputs, and how robust they are across models, tasks and decoding conditions.

02/PUBLICATIONS
6 PUB · 1 ACC · 5 REVIEW
Published
2026
CVPR

DASH: A meta-attack framework for synthesizing effective and stealthy adversarial examples

Nafi, A. A. N., Rahaman, H., Haider, Z., Mahfuz, T., Suya, F., Bhunia, S., & Chakraborty, P.
Adversarial ML / Meta-Attack
Published Paper ↗
2025
Nature Sci. Reports

A framework for mitigating malicious RLHF feedback in LLM training using consensus-based reward (COBRA)

Haider, Z., Rahman, M. H., Devabhaktuni, V., Moeykens, S., & Chakraborty, P.
Trustworthy ML / RLHF Integrity

ContributionConsensus-based reward that negates malicious human-in-the-loop RLHF feedback by aggregating across a mixed-trust expert pool rather than rejecting updates outright.

Published Paper ↗ Demo ↗
2025
IEEE Access

Intelligent Layer Sharing (ILASH): A predictive neural architecture search framework for multi-task applications

Rahman, M. H., Haider, Z., Rizvee, M. M., Shomaji, S., & Chakraborty, P.
NAS / Multi-Task Learning
Published Paper ↗
2025
Nature Sci. Reports

An automated multi-parameter neural architecture discovery framework using ChatGPT in the backend

Rahman, M. H., Haider, Z., & Chakraborty, P.
NAS / LLM Tooling
Published Paper ↗
2023
J. Microbiol. Methods

Towards digital diagnosis of malaria: How far have we reached?

Aqeel, S., Haider, Z., & Khan, W.
Medical AI / Review
Published Paper ↗
2022
IEEE IMPACT

Assistive stick for visually impaired people

Khan, D. A., Zamir, M. A., Umar, M. S., & Haider, Z.
Embedded AI / Accessibility
Published Paper ↗
Accepted
2026
IEEE CARS

ACT: Anchor-calibrated trust estimation when corrupted sources form the majority

Haider, Z., Rahman, M. H., Glatter, S., & Chakraborty, P.
Trustworthy ML / Trust Estimation

ContributionAnchor-calibrated trust estimation that recovers trusted and corrupted source cohorts when agreement-based methods become directionally ambiguous under corrupted majorities.

Accepted Demo ↗
2026
ACM CCS

(A)iSpy: Parasitic trojans for machine learning infrastructure

Rahaman, H., Xu, Q., Haider, Z., Chakraborty, P., Bhunia, S., & Suya, F.
ML Infrastructure Security
Accepted arXiv ↗
Under Review
—

BLADE: Bit-flip attacks for semantic misdirection in quantized image captioning

Haider, Z., Rahman, M. H., Moeykens, S., Devabhaktuni, V., & Chakraborty, P.
AI Model Security / Bit-Level VLM Faults

ContributionDifferentiable bit-level fault analysis for quantized vision–language models, studying controlled semantic steering under sparse parameter faults (BLIP, BLIP-2, GIT).

Under Review arXiv ↗ Demo ↗
—

SMASH: Probing speech recognition robustness via semantically targeted bit flips

Haider, Z., Rahman, M. H., Bumin, A., & Chakraborty, P.
AI Model Security / Targeted Speech Faults

ContributionProbes speech-recognition robustness via sparse, semantically targeted parameter bit-flips, characterizing controlled output changes and transferability.

Under Review
—

HARP: Measuring harm amplification in multi-agent LLM systems

Rahman, M. H., Haider, Z., Mahfuz, T., & Chakraborty, P.
Multi-Agent LLM Harm / Harm Amplification
Under Review arXiv ↗
—

Frontier AI releases should publish signed security manifests

Bumin, A., Haider, Z., & Chakraborty, P.
AI Policy / Position Paper
Under Review
03/PATENTS_AND_IP
04 FAMILIES
01
PCT

Methods and systems for the automated generation of neural network architectures

Chakraborty, P., Rahman, M. H., & Haider, Z.
WO 2025/166046 A1 · PCT/US2025/013847 · filed Jan 30, 2025 · priority US 63/627,621 (Jan 31, 2024)
Filed Patent ↗
02
PCT

Framework for mitigating malicious RLHF feedback in LLM training using consensus-based reward

Haider, Z., Rahman, M. H., Moeykens, S., & Chakraborty, P.
PCT application · filed Mar 9, 2026 · priority US 63/768,361
Filed UMaine Record ↗
03
US Prov.

Explainable agentic framework for determining cross-PDK migration difficulty

Chakraborty, P., Regan, T., Mahfuz, T., & Haider, Z.
US Provisional 64/075,062 · filed May 26, 2026
Filed
04
US Prov.

Bit-flip induced semantic drift in multimodal transformer models

Haider, Z., Rahman, M. H., Moeykens, S., & Chakraborty, P.
US Provisional 63/930,131
Filed
04/APPOINTMENTS

University of Maine

Aug 2022 – Present
  • Research Assistant, ECE May 2026 – Present
    AI security and trustworthy machine learning research.
  • Graduate Assistant / AI & ML Consultant, ARCSIM May 2023 – May 2026
    Advanced Research Computing, Security & Information Management.
  • Research & Teaching Assistant Aug 2022 – May 2023
    Split research and teaching appointment.

Aligarh Muslim University

  • Assistant Professor — ZHCET Jan 2015 – Aug 2022
  • Guest Faculty — ZHCET Aug 2013 – Feb 2014
  • Teaching Assistant — Computer Engineering (Data Structure, Minor Project, Software labs) Aug 2011 – Jul 2013
05/FUNDING_AND_FELLOWSHIPS

Funding

SentriAlign: Training-Signal Integrity for LLMs

Role
Co-Investigator
Program
MIRTA 8.0 Accelerator
Amount
$25,000
Term
Jan 14 – May 11, 2026

Commercialization of the COBRA consensus-reward line of work.

Honors & Fellowships

  • Maulana Azad National Fellowship (MANF), Government of India 2014
  • Graduate Aptitude Test in Engineering (GATE) Scholarship, Government of India 2011–13
  • Teaching & Research Assistantship, TEQIP-II
  • Graduate Assistantship, University of Maine 2022–
06/ENGINEERING_SYSTEMS
03 SYSTEMS
Human Evaluation & AI Assessment Platform2026
Streamlit · Supabase · Python

Private blinded multi-rater evaluation platform built for human validation of LLM-as-judge metrics in multi-agent AI-security experiments, with controlled rater assignment, randomized presentation, autosave/resume, structured annotation, and reproducible exports.

Adversarial Evaluation Tooling2025
Internal Systems / Streamlit

Full-stack experimental tooling for structured review and downstream analysis of adversarial studies — repeatable AI-evaluation workflows via a Streamlit interface and Python export pipelines: structured scoring, diff-based review, and CSV export.

Distributed / Edge LLM Inference2024
Heterogeneous NVIDIA Hardware

Resource-aware LLM execution across heterogeneous NVIDIA hardware, partitioning models layer-by-layer across Jetson Orin Nano, Xavier NX, AGX Orin and an RTX A6000, with models up to the 15B scale (StarCoder).

07/TALKS_AND_PRESENTATIONS

Invited Talks & Guest Lectures

  • Bit-Flip Attacks: PBS, GenBFA, and BLADE — UMaine, ECE 479/579 Apr 3, 2026
  • Data Poisoning, RLHF, and COBRA — UMaine, ECE 479/579 Mar 30, 2026
  • Large Language Models — Purdue University Northwest, ECE seminar Mar 5, 2026
  • Artificial Intelligence Seminar — UMaine Apr 2025
  • AI and its use cases — CNLU, Bihar, India Apr 2025
  • AI Security — UMaine Mar 2024
  • LLMs and Security — CNLU, India 2024
  • Invited by UGC-HRDC to speak on MOOCs

Conference Presentations, Posters & Demos

  • BLADE: Bit-level Fault Analysis — Poster, Warren B. Nelms Annual IoT Conf., Gainesville, FL Dec 2025
  • Enhancing RLHF Robustness with Multi-Agent Consensus — Demo, Nelms IoT Conf. Dec 2025
  • Mitigating Malicious RLHF Feedback (COBRA) — Sandia National Labs Annual ML/DL Workshop Aug 2025
  • COBRA — WHiSPerS, University of New Hampshire Mar 2025
  • North Atlantic Right Whale Calls Monitoring — 184th Meeting of the Acoustical Society of America, Chicago May 2023

Outreach & Training

  • AI-Driven Tools for Decision-Making in Forest Management — Co-presenter, SELF 3.0 Summer Camp, UMaine Jun 26, 2026
  • Trainer for lecturers of Sayid Hamid Senior Secondary School on Google Tech
  • Delivered educational lectures for the SWAYAM course
08/TEACHING
200+ STUDENTS / YR
  • ECE275 Sequential Logic Lab — UMaine (TA)
  • ECE473 Microprocessor Lab — UMaine (TA)
  • COC4010 Information Security — 2020–2021
  • COA3600 Data Structures — 2020–2022
  • COE0431 Internet Tools (networking, sockets) — 2015–2019
  • COA1910 C Programming — 2015–2022
  • COA446N Selected Topics in CE-I (OS, DS) — 2015–2019
  • COA447N Selected Topics in CE-II (Networking, DB) — 2013
  • COC0203 Object-Oriented Programming — 2013
09/MENTORING
  • 2019 Human Activity Recognition using smartphone sensors
  • 2019 Assistive Stick for visually impaired people Also a publication
  • 2018 Disease Risk Prediction (IHDPS)
  • 2017 2π Surveillance System
  • 2016 Network Report Engine
  • 2016 Hajj Guide via Offline GPS
  • 2013 A framework for real-time online auctioning
  • 2012 Dynamic News Alert Android application
  • 2012 SyncWorld — file storage & synchronization
  • 2012 Chatting Website
10/SERVICE_AND_PEER_REVIEW
  • Reviewer — Nature Scientific Reports
  • Reviewer — IEEE UPCON 2019 (two conference papers)
  • Judge — 2026 University of Maine Student Symposium Apr 17, 2026
Earlier Academic Service (AMU) 13 roles
Incharge
Interactive Instructional Resource Unit · Computer Programming Lab · MIS Cell
Coordinator
GATE Coaching · EdX Campus · Coursera Campus
Convener
MOOCs (AMU) · OGRM
Faculty Advisor
AMU-OSS · CodeChef ZHCET · Google Developer Student Club
Coordinator
ZHCET-NA Cell · Warden, Sir Ross Masood Hall
Committees
AICTE · NBA · NAAC · Counseling · Discipline · ERP
11/TECHNICAL_SKILLS
Languages
Python, C/C++, SQL, Shell, Git, Jupyter, LaTeX
Frameworks
PyTorch, scikit-learn, NumPy, CuPy, Pandas, Streamlit, REST APIs
AI / Security
Adversarial ML, bit-flip & fault injection, RLHF & reward modeling, LLM & agentic evaluation
Systems / Edge
Slurm, Apptainer, NVIDIA containers, multi-GPU/HPC, GCP, Jetson (Orin Nano, Xavier NX, AGX)
Broader Technical History
Languages
Java, JavaScript, PHP, Prolog, LISP
Compiler
Lex, YACC
Databases
Oracle, MySQL, SQLite
Hardware
Raspberry Pi, Verilog, VCU118 FPGA, Vivado, Vitis, Quartus
Tools/IDEs
NetBeans, Eclipse, VSCode, Google Colab, Kaggle, Cursor
12/EARLIER_RESEARCH
  • 2023 Game graphics for a custom OS — Pong in user space via PS/2 & kernel integration
  • 2022 N. Atlantic right-whale call classification — self-supervised DL under class imbalance
  • 2022 BEiTv2 masked image modeling for breast-cancer imaging
  • 2022 Human Activity Recognition via MS Kinect / smartphone sensors
  • 2022 Assistive stick for the visually impaired (object detection + audio)
  • — Healthcare data science — blood-smear cell detection · exploratory project, halted
  • 2012 Extension to Lex & YACC — PASCAL-like authoring interface
  • 2012 Scheme for normalization of metrics
13/PROFESSIONAL_DEVELOPMENT
Short Courses & Workshops selected
  • 2021 Industrial Training on Automation — TSC, NSIC
  • 2020 Moodle LMS FDP — AMU & IIT Bombay
  • 2020 Online Teaching — UGC-HRDC
  • 2018 Deep Learning STC (TEQIP-III) — ZHCET, AMU
  • 2018 Special Summer/Winter School — UGC-HRDC, AMU
  • 2015 Cyber Security: Fundamentals & Advancements — IIT-D (TEQIP-II)
14/EDUCATION
  • Ph.D., Electrical & Computer Engineering — University of Maine, Orono, ME, USA (in progress) Aug 2022 – Present
  • M.Tech., Computer Engineering — Aligarh Muslim University, India (GPA 9.33/10, Hons.) Aug 2011 – Jul 2013
  • B.Tech., Computer Science & Engineering — UP Technical University (now AKTU), India (76.2%, Hons.) Aug 2007 – Jul 2011