BS CS @ FAST-NUCES ('27) · Open to Research & AI Engineering Collaborations

AI/ML Engineer in Training · Researcher · Applied Systems Builder

Imaad Fazal

Building interpretable, resource-aware AI systems for high-stakes environments.

I design, audit, and engineer machine learning pipelines across clinical prediction, aviation safety, computer vision, and agentic workflows—grounding every model in physical constraints, honest evaluation, and human-interpretable telemetry.

Explore Selected Work View Résumé

GitHub LinkedIn

BASE STATION

Islamabad, PK

INSTITUTION

FAST-NUCES (’27)

PRIMARY FOCUS

Clinical · Aviation · XAI

STATIC SCHEMATIC MODE // REDUCED MOTION READYCLICK STATION TO JUMP

CREDIBILITY CONSTELLATION // VERIFIED TELEMETRYSELECT SIGNAL TO INSPECT EVIDENCE

Top 10 / 575+→ WORK

IEEE IES GenAI Challenge 2026

Global finalist shortlisted for IRAI Conference, Melbourne

First Author→ RESEARCH

Aviation Safety AI Research

Physics-constrained ADS-B & sub-50 ms decision agents

0.927 AUROC→ WORK

SepsisAI Clinical Pipeline

~24-min lead time & 8.9× risk separation on MIMIC

HackUPC ’26→ CREDENTIALS

Barcelona International Build

Represented Pakistan & FAST-NUCES in Skyscanner challenge

3 AI Labs→ EXPERIENCE

TechGIS · MIDL/NCAI · NASTP

Applied ML across healthcare, aerospace & agentic systems

Band 8.0 / 9.0→ CREDENTIALS

IELTS Academic & MIT CITI 100%

Credentialed for international research & MIMIC-IV access

OBS-00 // MISSION BRIEFResearch Identity & Systems Philosophy

Bridging rigorous empirical evaluation with deployable real-time tooling where false confidence has a real cost.

As a Computer Science undergraduate at FAST-NUCES Islamabad (expected graduation June 2027), my work focuses on turning experimental machine learning into reliable, interpretable systems. Whether modeling ICU sepsis trajectories on MIMIC clinical time-series or auditing ADS-B aircraft approach telemetry, I prioritize honest baselines, leakage-free validation, and post-hoc explainability.

Across internships at TechGIS, the National Centre for Artificial Intelligence (MIDL/NCAI), and the National Aerospace Science and Technology Park (NASTP), I have worked across the full stack of modern applied AI: from deep physiological signal modeling and hybrid Vision-Language OCR pipelines to sovereign/on-prem multi-agent architectures and sub-50 ms decision dashboards.

Physics-Constrained & Aviation Safety ML

First Author · IEEE IES GenAI Challenge 2026 Top 10 / 575+ · Safety Science submission

Coupling generative trajectory augmentation (cVAE) with hard aerodynamic feasibility filters and SHAP explainability for unstable aircraft approach and runway excursion risk modeling.

Clinical Time-Series & Medical Imaging AI

SepsisAI (0.927 AUROC, ~24m lead time) · MIT/CITI Credentialed Researcher

Developing early-warning clinical pipelines on MIMIC/PhysioNet datasets, ECG temporal convolutional networks, and Grad-CAM visual attention workflows for transparent decision support.

Agentic Systems, VLM OCR & Edge Inference

TechGIS & MIDL/NCAI Internships · HackUPC 2026 Barcelona

Architecting multi-agent reasoning, LangChain tool-calling clinical workflows, Nanonets-OCR + Ollama pipelines, and resource-aware inference for sovereign/air-gapped environments.

ACADEMIC TELEMETRY

FAST-NUCES Islamabad

National University of Computer and Emerging Sciences (FAST-NUCES)

Degree ProgramBachelor of Science in Computer Science

Expected GraduationJune 2027

CampusIslamabad, Pakistan

Relevant Coursework & Foundations

Machine LearningArtificial IntelligenceData Structures & AlgorithmsProbability & StatisticsLinear AlgebraData MiningMultivariable CalculusPython ProgrammingOperating SystemsCompiler ConstructionDatabase SystemsHigh Performance Computing

Evaluation Integrity Principle: Every metric reported in this observatory distinguishes between retrospective research prototypes, leakage-corrected baselines, and submitted manuscripts.

OBS-01 // FLAGSHIP STATIONSSelected Research & Engineering Case Studies

Systems engineered for interpretability, physical constraints, and real-world telemetry.

Select any research station below to inspect its full problem formulation, system architecture, leakage-audited metrics, and evaluation scope.

STN-01 // FLIGHT DECKSubmitted to IEEE/IRAI 2026 · Top 10 / 575+ Challenge Finalist

ADS-B GLIDESLOPE TELEMETRY // 3.0° REFcVAE + PHYSICS FILTER

10 NM APPROACH GATETHRESHOLD // <50MS AGENT ADVISORY

Research & Aviation Safety · 2026 · First Author & Lead System Architect

Early Detection of Unstable Aircraft Approaches Using GenAI & Autonomous Decision Agents

Physics-Constrained ADS-B Trajectory Modeling & Sub-50 ms Decision Advisory System

First-author research submission (Top 10 from 575+ global submissions in the IEEE IES Generative AI Challenge 2026) combining leakage-corrected XGBoost baselines, physics-filtered Conditional VAE augmentation, SHAP explainability, and a FastAPI autonomous decision-agent dashboard.

Top 10 / 575+

Challenge Standing

0.8962

Physics-cVAE AUROC

0.8923

Honest Baseline AUROC

PythonXGBoostPyTorch (cVAE)SHAP

STN-02 // CLINICAL ICUResearch Prototype · Verified Public Repository

MIMIC ICU // HEMODYNAMIC TRAJECTORYAUROC 0.927 · ~24M LEAD

HR · MAP · LACTATE · WBCRESEARCH EVALUATION ONLY

Healthcare AI · Jan 2026 · Lead ML Researcher & Pipeline Engineer

SepsisAI — Early Sepsis Prediction on Clinical Time-Series

ICU Early-Warning Pipeline on MIMIC Physiological & Laboratory Trajectories

End-to-end clinical time-series machine learning pipeline on the MIMIC dataset achieving 0.927 AUROC with an approximate 24-minute early-warning lead time and 8.9× risk separation across patient cohorts.

0.927

Retrospective AUROC

~24 min

Early Warning Lead

8.9×

Risk Separation

PythonPyTorchScikit-LearnPandas

Repo

STN-03 // BARCELONAShipped at HackUPC 2026 (Barcelona) · Skyscanner Challenge Entry

DreamTrip — Emotion-Aware AI Travel Discovery Platform — Barcelona geodesic route discovery illustration

BCN // HACKUPC 2026MULTI-CITY SENTIMENT EMBEDDINGS

Agentic AI · Apr 2026 · AI/ML & Backend Workflow Engineer

DreamTrip — Emotion-Aware AI Travel Discovery Platform

Sentiment-Driven Multi-City Embeddings & Voice-Narrated Discovery (Skyscanner Challenge)

Built in under 36 hours at HackUPC 2026 in Barcelona (representing Pakistan and FAST-NUCES). Converts emotion- and intent-rich natural language prompts into ranked destination matches with Gemini reasoning and ElevenLabs narration.

< 36 hrs

Build Window

Multi-City

Core Engine

Text + Voice

Modalities

PythonGemini APIElevenLabsFastAPI

Repo

STN-04 // MIDL-NCAILab Research & Engineering Workflow · Completed at NCAI

MIDL / NCAI // HYBRID VLM + AGENT WORKFLOWSYNTHETIC TRACE

01. OCR + VLM

Nanonets + Ollama

Handwritten layout parse

02. TOOL AGENT

LangChain Lookup

Dosage & side-effects

03. CLINICAL Q&A

MedGemMA 4.0

Grounded synthesis

NON-PRESCRIBING RESEARCH PIPELINENCAI ISLAMABAD

Healthcare AI · Jul – Sep 2025 · Artificial Intelligence Intern (MIDL — NCAI)

Medical Prescription Intelligence & Agentic Clinical Workflow

Hybrid VLM Prescription OCR + LangChain Tool-Calling Agents + MedGemMA 4.0

Developed during the AI Internship at the Medical Imaging and Diagnostics Lab (MIDL / NCAI). Integrates Nanonets-OCR, Ollama VLMs, LangChain tool-calling agents, and MedGemMA 4.0 for structured prescription extraction and medication safety lookup.

Hybrid VLM

OCR Architecture

LangChain

Agent Framework

MedGemMA 4.0

Clinical LLM

PythonNanonets-OCROllamaVision-Language Models

STN-05 // VISION LABResearch Prototype · Verified Public Repository

RESNET18 // GRAD-CAM SALIENCY AUDITPYTORCH VISION

SPATIAL ATTRIBUTION LAYERRESEARCH PROTOTYPE

Computer Vision · Jun 2025 · ML Developer & Researcher

Interpretable Pneumonia Detection from Chest X-Rays

ResNet18 Thoracic Radiograph Classifier with Grad-CAM Spatial Attention

Deep learning medical imaging pipeline in PyTorch pairing a fine-tuned ResNet18 classifier with Gradient-weighted Class Activation Mapping (Grad-CAM) to expose spatial regions driving thoracic radiograph predictions.

ResNet18

Backbone

Grad-CAM

Explainability

Thoracic CXR

Domain

PythonPyTorchResNet18Grad-CAM

Repo

OBS-02 // RESEARCH LOGManuscript Submissions & Empirical Evaluation

Research outputs with transparent methodology and explicit publication status.

Distinguishing submitted manuscripts, global competition recognition, and ongoing credentialed clinical dataset investigations.

RES-01 // AVIATION-GENAI· 2026 · First Author (Imaad Fazal et al.)

Submitted · Top 10 Global Finalist

Early Detection of Unstable Aircraft Approaches Using GenAI & Autonomous Decision Agents

IEEE IES Generative AI Challenge 2026 — IRAI Conference, Melbourne

Presents a physics-constrained ADS-B approach analysis pipeline and real-time decision-agent architecture. Audits and resolves chronological window leakage (correcting an initial 0.9513 figure to a rigorous 0.8923 honest XGBoost baseline) before applying Conditional VAE trajectory synthesis with hard aerodynamic filtering to reach 0.8962 AUROC.

[1]Shortlisted in the Top 10 out of 575+ global submissions in the IEEE IES Generative AI Challenge 2026.
[2]Explicit chronological data-leakage audit: documented the 0.9513 → 0.8923 baseline correction prior to generative augmentation.
[3]Hard physics filtering on cVAE synthetic trajectories enforcing vertical speed, glide-slope, and groundspeed constraints.
[4]Sub-50 ms FastAPI + XGBoost + TreeSHAP + LLM Autonomous Decision Agent dashboard for interpretable approach advisories.

Status: Submitted manuscript & Top 10 Challenge Finalist associated with IEEE/IRAI 2026. Evaluated on historical/simulated ADS-B telemetry; not claimed as an already-published journal article or operational avionics deployment.

Google Scholar Profile

RES-02 // RUNWAY-RISK· 2026 · Research Manuscript Submission

Submitted Manuscript

Predicting Runway Risk Excursion

Submitted to Elsevier Safety Science

Investigates predictive risk modeling for aircraft runway excursion precursors during approach and landing rollout phases, emphasizing interpretable risk factors and safety-critical evaluation discipline.

[1]Submitted to Elsevier Safety Science in 2026.
[2]Focuses on quantifiable precursor modeling and actionable safety telemetry for runway excursion mitigation.
[3]Complements the ADS-B unstable approach framework within a unified aviation safety research agenda.

Status: Manuscript submitted to Elsevier Safety Science (under review / submitted status; not presented as accepted or published).

Google Scholar Profile

RES-03 // CLINICAL-ML· 2025 – 2026 · Credentialed Researcher (MIT / CITI Program 100% Completion)

Active Clinical ML Research

Clinical Time-Series & Physiological Signal Modeling on MIMIC-IV / PhysioNet

PhysioNet / MIT MIMIC-IV Credentialed Clinical ML Research

Ongoing empirical research developing and evaluating machine learning pipelines on de-identified intensive-care and physiological waveform datasets, spanning ICU sepsis early warning (SepsisAI) and multi-scale ECG anomaly detection.

[1]Completed MIT / CITI Data or Specimens Only Research training (100% score, Jan 2026) for authorized MIMIC-IV clinical database research.
[2]Developed SepsisAI achieving 0.927 retrospective AUROC, ~24-minute lead time, and 8.9× risk separation on MIMIC clinical time-series.
[3]Explored Temporal Convolutional Networks (TCNs) and SHAP-explained biomedical voice/signal models across PhysioNet cohorts.

Status: Ongoing research activity and credentialed dataset evaluation (no fabricated paper title, DOI, or clinical trial claim; MIT/CITI credential represents research training and database authorization, not MIT employment).

OBS-03 // FIELD TELEMETRYApplied AI & Research Internships

Hands-on engineering across national AI labs, aerospace technology parks, and industry R&D.

Reverse-chronological record of technical internships in Islamabad and Rawalpindi.

EXP-01 // TECHGISCompleted

Applied AI/ML Intern

TechGIS (Pvt.) Limited15 July 2026 – 15 August 2026Islamabad, Pakistan

Engineered and evaluated agentic workflows, semantic retrieval pipelines, and resource-aware inference architectures for production-oriented and sovereign AI deployments.

▹Worked across agentic AI, multi-agent reasoning, RAG, semantic retrieval, vector embeddings, generative/cognitive AI, and production-oriented AI engineering.
▹Explored sovereign, air-gapped, and on-premise AI deployments alongside computer vision, geospatial AI, edge AI, efficient model inference, and neural architecture evaluation.
▹Contributed to research, experimentation, benchmark evaluation, and deployment-oriented workflows under Manager Hameem Ahmad and Supervisor Inayat Arshad.

Mentored by Manager Hameem Ahmad & Supervisor Inayat Arshad.

Agentic AIMulti-Agent ReasoningRAG & EmbeddingsSovereign / Air-Gapped AIGeospatial & Edge AI

EXP-02 // MIDL-NCAICompleted

Artificial Intelligence Intern

Medical Imaging and Diagnostics Lab (MIDL) — National Centre for Artificial Intelligence (NCAI)July 2025 – September 2025Islamabad, Pakistan

Researched and built hybrid Vision-Language prescription OCR pipelines and tool-calling clinical information agents at Pakistan's flagship national AI research lab.

▹Advanced a medical prescription OCR pipeline combining classical document processing and modern Vision-Language Model (VLM) techniques, integrating Nanonets-OCR and Ollama.
▹Built LangChain agents with deterministic tool-calling for medication instruction and side-effect retrieval, incorporating MedGemMA 4.0 for grounded clinical Q&A.
▹Recognized by lab leadership for technical excellence, professionalism, and effective cross-functional teamwork.

Recognized for technical excellence, professionalism, and cross-functional teamwork.

Vision-Language ModelsNanonets-OCR & OllamaLangChain Tool-CallingMedGemMA 4.0Healthcare AI

EXP-03 // NASTPCompleted

AI & ML Intern

NASTP — National Aerospace Science and Technology ParkAugust 2024 – October 2024Rawalpindi, Pakistan

Collaborated on deep learning model training, physiological signal analysis pipelines, and software quality engineering within an aerospace and defense technology ecosystem.

▹Collaborated on AI model training, data visualization and analysis pipelines, and ECG anomaly detection using deep learning (Temporal Convolutional Networks).
▹Participated in structured code reviews and improved pipeline modularity, execution performance, and agile engineering practices.

Deep LearningECG Signal Anomaly DetectionAnalysis PipelinesCode Quality & Agile

OBS-04 // SYSTEM ARCHIVEComplete Engineering & Research Index

Filterable archive across clinical ML, computer vision, hackathons, and systems programming.

STN-01 // FLIGHT DECK2026

Early Detection of Unstable Aircraft Approaches Using GenAI & Autonomous Decision Agents

Submitted to IEEE/IRAI 2026 · Top 10 / 575+ Challenge Finalist

PythonXGBoostPyTorch (cVAE)SHAP

STN-02 // CLINICAL ICUJan 2026

SepsisAI — Early Sepsis Prediction on Clinical Time-Series

Research Prototype · Verified Public Repository

PythonPyTorchScikit-LearnPandas

GitHub

STN-03 // BARCELONAApr 2026

DreamTrip — Emotion-Aware AI Travel Discovery Platform

Shipped at HackUPC 2026 (Barcelona) · Skyscanner Challenge Entry

PythonGemini APIElevenLabsFastAPI

GitHub

STN-04 // MIDL-NCAIJul – Sep 2025

Medical Prescription Intelligence & Agentic Clinical Workflow

Lab Research & Engineering Workflow · Completed at NCAI

PythonNanonets-OCROllamaVision-Language Models

STN-05 // VISION LABJun 2025

Interpretable Pneumonia Detection from Chest X-Rays

Research Prototype · Verified Public Repository

PythonPyTorchResNet18Grad-CAM

GitHub

ARC-01 // VOICE BIOMARKERSDec 2025

Parkinson’s Disease Early Detection via Biomedical Voice Telemetry

Completed Research Build · Verified Public Repo

Biomedical voice ML pipeline comparing classification architectures on vocal fundamental frequency jitter, shimmer, and harmonic-to-noise ratio biomarkers, paired with SHAP feature attributions for medical interpretability.

PythonXGBoostScikit-LearnSHAP

GitHub

ARC-02 // CARDIAC TCNAug 2024

ECG Anomaly Detection with Temporal Convolutional Networks

Completed Pipeline · Verified Public Repo

Deep physiological signal classification pipeline using Temporal Convolutional Networks (TCNs) on the ECG-5000 dataset to detect cardiac waveform anomalies at scale.

PythonTemporal Convolutional Networks (TCN)Deep LearningNumPy

GitHub

ARC-03 // HEATSHIELD AI2026

FortyGuard HeatShield AI — Urban Heat-Risk Radar Prototype

Hackathon Prototype Built (FortyGuard ’26)

Developed an AI-driven heat-risk radar prototype through the FortyGuard Hackathon ’26, modeling urban thermal vulnerability and actionable heat-safety advisories.

PythonGeospatial AIThermal Risk ModelingRadar Visualization

ARC-04 // AUDIO RECJun 2025

Smart Music Recommender

Completed & Deployed via Gradio · Verified Public Repo

Collaborative and content-aware music recommendation engine operating on acoustic feature vectors and genre patterns, deployed with an interactive Gradio interface for real-time personalized suggestions.

PythonScikit-LearnGradioPandas

GitHub

ARC-05 // C++ ENGINEAug 2024

Candy Crush Clone (Academic Team-Lead Project)

Completed Academic Lead Project · Verified Public Repo

Led an academic engineering team building a feature-complete match-3 game in C++ and SFML with cascading board physics, score combos, cross-platform compatibility, and a polished UI.

C++SFML LibraryOOP ArchitectureState Machines

GitHub

OBS-05 // TECHNICAL ATLASCapabilities by Engineering Domain

Structured domain clusters grounded in real research pipelines—no arbitrary percentage bars.

Every tool and framework listed below maps directly to coursework, lab internships, or verified repository builds.

DOM-01 // CORE ML & XAI

Machine Learning, Deep Learning & Explainability

Predictive modeling, neural architectures, and post-hoc interpretability

PyTorch

cVAE, ResNet18, Time-Series

TensorFlow

Deep Learning Pipelines

Scikit-Learn

Baselines & Calibration

XGBoost

Leakage-Audited Tabular & Time-Series

LightGBM

Gradient Boosted Ensembles

Hugging Face Transformers

Sequence & Multimodal

SHAP

TreeSHAP & Clinical Attribution

Grad-CAM

Radiograph Spatial Saliency

Applied across SepsisAI (MIMIC), ADS-B cVAE + XGBoost, Chest X-Ray ResNet18, and ECG TCN pipelines.

DOM-02 // GENAI & AGENTS

Generative AI, VLMs & Agentic Workflows

Tool-calling agents, multimodal OCR, semantic retrieval, and real-time APIs

LangChain Agents

Tool-Calling & Orchestration

RAG & Semantic Retrieval

Grounded Knowledge Pipelines

Vector Embeddings

Multi-City & Clinical Retrieval

Gemini API

Structured Reasoning & DreamTrip

Claude API

Agentic Evaluation Workflows

Ollama

Local / On-Prem VLM & LLM Inference

Nanonets-OCR

Hybrid Medical Document OCR

MedGemMA 4.0

Clinical Q&A Synthesis

ElevenLabs

Low-Latency Voice Synthesis

FastAPI

Sub-50 ms Inference & Agent APIs

Applied at TechGIS, MIDL/NCAI (prescription intelligence), HackUPC 2026 (DreamTrip), and IEEE IRAI advisory agents.

DOM-03 // RESEARCH FOCUS

High-Stakes Research & Application Domains

Physical constraints, clinical rigor, and resource-aware deployment

Healthcare AI & Clinical ML

ICU Sepsis & Physiological Signals

Aviation Safety ML

ADS-B Unstable Approach & Runway Risk

Trustworthy & Explainable ML

Leakage Audits, SHAP & Grad-CAM

Computer Vision

Medical Imaging & VLM Document Parsing

Agentic & Multi-Agent AI

Tool-Calling & Autonomous Advisories

Edge & Resource-Constrained AI

Sovereign / Air-Gapped & Low-Latency

Geospatial AI

TechGIS & Thermal Risk Radar

Core research themes spanning IEEE/IRAI, Elsevier Safety Science, MIT/CITI MIMIC-IV, and sovereign AI internships.

DOM-04 // LANGUAGES

Programming & Scientific Computing Languages

Systems programming, numerical computing, and full-stack implementation

Python

Primary ML, Research & FastAPI Stack

C++

Systems, Algorithms & SFML Engine

Operating Systems & Low-Level

SQL

Relational Querying & Cohort Extraction

Java

Object-Oriented Engineering

MATLAB

Numerical & Signal Analysis

HTML / CSS

Web Interfaces & Dashboards

PHP

Backend Web Scripting

Used across research codebases, FAST-NUCES systems coursework, and native C++ engines.

DOM-05 // DATA & INFRA

Data Engineering, Databases & Interactive Tooling

Analysis pipelines, relational systems, and rapid ML prototyping

Pandas & NumPy

Time-Series & Vectorized Processing

Matplotlib

Scientific & Telemetry Plotting

Jupyter

Reproducible Research Notebooks

Gradio & Streamlit

Interactive Model Demos

Git & GitHub

Version Control & Code Review

MySQL / SQL Server / Oracle

Relational Database Systems

VS Code & PyCharm

Development Environments

Bootstrap

Responsive UI Prototyping

Daily engineering stack for dataset curation, reproducible experiments, and interactive research demos.

OBS-06 // HONOURS & DOSSIERAwards, Credentials & International Engagement

Verified honours, research ethics credentials, and transparent hackathon status tracking.

Every international event entry explicitly distinguishes between in-person attendance, prototype builds, upcoming selections, and invitations.

01 // Honours, Research Authorizations & Certifications

May 2026Top 10 / 575+

IEEE IES Generative AI Challenge 2026 — Top 10 Global Finalist

IEEE Industrial Electronics Society / IRAI Conference, Melbourne

Shortlisted in the global Top 10 from 575+ submissions as First Author for 'Early Detection of Unstable Aircraft Approaches Using GenAI & Autonomous Decision Agents'.

December 2024Honorable Mention

ICPC 2024 — Honorable Mention

International Collegiate Programming Contest

Recognized in the International Collegiate Programming Contest (ICPC) for algorithmic problem-solving.

January 2026100% Score

CITI Program — Data or Specimens Only Research Training

MIT / PhysioNet MIMIC-IV Research Credential

Completed mandatory human-subjects data ethics and governance training with a 100% score, authorizing research access to the MIT MIMIC-IV clinical database.

December 2025Graduate

McKinsey.org Forward Program — Graduate Credential

McKinsey & Company (McKinsey.org)

Completed structured problem-solving, analytical communication, and digital leadership curriculum.

September 2026Overall Band 8.0 / 9.0

IELTS Academic — English Proficiency

IELTS Official Assessment

C1/C2 academic English proficiency credential supporting international research and graduate study.

July 2025C1 Advanced (62/100)

EF SET English Certificate — C1 Advanced

EF Standard English Test

Certified C1 Advanced English reading and listening proficiency.

December 2024Certified

Google Soft Skills Program — Certificate

Google

Professional communication and collaborative engineering program completion.

02 // International Hackathon & Global Selection Ledger (Status-Audited)

Explicit Distinction: Attended vs. Prototype Built vs. Selected/Invited

April 2026 · Universitat Politècnica de Catalunya · Barcelona, SpainAttended & Built On-Site

HackUPC 2026

Represented Pakistan and FAST-NUCES in person in Barcelona; built DreamTrip (emotion-aware AI travel recommender with Gemini API, ElevenLabs, and multi-city embeddings) for the Skyscanner challenge in under 36 hours.

2026 · International HackathonPrototype Built

FortyGuard Hackathon ’26 (HeatShield AI)

Developed FortyGuard HeatShield AI, an AI-driven urban heat-risk radar prototype.

March 2026 · Cognitive Abilities TrackBenchmark Entry Submitted

Google DeepMind × Kaggle — Measuring Progress Toward AGI

Submitted a cognitive-abilities benchmark and evaluation entry to the Google DeepMind × Kaggle challenge.

Late 2026 (Upcoming) · FinlandSelected / Invited (Upcoming)

Junction 2026

Selected for participation in Junction 2026 Finland (upcoming as of October 2026; status configurable in src/data/achievements.ts).

2026 · Lisbon & TokyoSelected / Invited

ETHGlobal Lisbon & ETHGlobal Tokyo

Selected/accepted for participation in ETHGlobal Lisbon and ETHGlobal Tokyo (listed as selection/invitation only, not claimed as attended).

2026 · University of MichiganSelected (Could Not Attend)

MHacks 2026

Selected as a participant for MHacks 2026; could not attend in person.

2026 · Global / Hybrid CohortsSelected / Invited

Additional 2026 International Selections & Civic AI Programs

Selected/accepted across AI for Civic Innovation (Code for Pakistan / Ctrl+Aid), Hack-Nation Global AI Hackathon, Hack Dearborn: Conjure Reality, DeepTech GigaHack, Knight Hacks, and HackUMBC.

03 // Leadership, Community Service & Technical Engagement

Research & Planning Team Member, Meri Kahani — FAST Outreach & Engagement Society, Islamabad.
Community Service: Mentored and guided pre-university students preparing for engineering and CS entrance examinations.
Healthcare AI Chatbot Assistant — recognized/mentioned by Rise & Recognize, Dubai.
Technical Conferences & Workshops: Agentic AI Masterclass (AI Data Yard), Indus AI Week 2026 (AI Data Yard), Indus Hospital Healthcare Conference, AI Business Meetup (NICAT/NASTP), EduNautics Workshop (NSTP).
Extracurricular: Member, FAST Futsal Team — competed in NASCON (National Speed Contest).

OBS-07 // SIGNAL LINKDirect Research & Engineering Contact

Let’s build AI systems that earn trust.

Available for AI/ML research collaborations, clinical or aviation safety ML projects, graduate research inquiries, and applied AI engineering opportunities.

PRIMARY TRANSMISSION CHANNEL

imdufazal@gmail.com

Direct email is the fastest way to reach me regarding research supervision, applied ML internships, hackathon builds, or technical collaborations.

Compose EmailOpen / Download Résumé (PDF)

LOCATION: ISLAMABAD, PAKISTAN (UTC+5)FAST-NUCES BS CS ’27

in/imaad-fazal

GitHub

github.com/Fastian-afk

Google Scholar

Citations & Submissions

Devpost

devpost.com/imdufazal

Kaggle

kaggle.com/imaadfazal