Exploring PhD research opportunities

RESEARCHER. ENGINEER. BUILDER.

Making AI agents remember better.

I'm Ali Ahmad Qureshi, an MS Computer Science researcher investigating long-term memory, retrieval, and reliability in LLM-based agents.

I bring research and software engineering together, from studying retrieval failures to building AI systems that run in real environments.

Explore my research Download CV

LinkedIn ↗GitHub ↗ORCID ↗

3.91 / 4.00MS CGPA

02Submitted manuscripts

2019NUST Rector's Gold Medal

2027Expected MS graduation

01 / RESEARCH

Questions I'm working on.

I study how memory and retrieval systems can make autonomous language-model agents more dependable over long interactions.

FEATURED MANUSCRIPTSubmitted · October 2026

EA-MEM: Failure-Tolerant Retrieval and Maintenance for LLM Agent Memory

Submitted to Complex & Intelligent Systems · Springer Nature

EA-MEM investigates failures in the way LLM agents extract, retrieve, and maintain memories. I extended the A-MEM framework with failure-tolerant metadata handling, dense and BM25 hybrid retrieval, graph-aware reranking, importance-based forgetting, and memory consolidation.

The work evaluates both retrieval quality and reliability under targeted failures, with a focus on preserving useful information as an agent's memory grows.

Agent memoryHybrid retrievalBM25Failure testing

CLOUD COMPUTING RESEARCHSubmitted · June 2026

A Priority-Aware Load-Normalized VM Scheduling Algorithm for Efficient Cloud Task Allocation

Research on priority-aware cloud task scheduling, balancing virtual-machine workload with task importance to improve resource allocation and scheduling efficiency.

Submitted to SN Computer Science · Springer Nature

02 / SELECTED PROJECTS

Ideas turned into working systems.

A selection of research prototypes, local AI applications, and engineering tools.

LLM AGENTS · MEMORY RESEARCH

EA-MEM — Agent Memory Framework

Developed a research framework for reliable memory retrieval and maintenance in LLM agents, combining hybrid search, failure handling, and memory consolidation.

ON-PREMISES · RETRIEVAL-AUGMENTED GENERATION

Local RAG Knowledge Assistant

Built an offline-first chatbot for organizational documents with semantic retrieval, incremental indexing, local LLM inference, and a desktop interface.

FPGA / SOC · DEVELOPMENT AUTOMATION

GENYSYS — Generate Your System

Developed a web-based platform for peripheral configuration, address mapping, HDL generation, pin assignments, Vivado builds, and artifact delivery.

CLOUD COMPUTING · SCHEDULING

Priority-Aware VM Scheduling

Investigated how load-normalized scheduling and task priorities can improve workload distribution and resource allocation across cloud virtual machines.

★ RECTOR'S GOLD MEDAL

FPGA-Based Sign Language Recognition

Implemented a resource-aware convolutional neural network on Xilinx Zynq-7000 hardware using Vivado HLS and Verilog, combining machine learning with hardware design.

Employer projects are described at a high level. Confidential code and documents are not publicly shared.

03 / EXPERIENCE

Engineering, beyond the prototype.

Experience delivering software, automation, and AI applications in practical environments.

From embedded software to AI systems.

My professional work connects low-level engineering with full-stack software and applied AI.

View full CV ↗

Assistant Manager (Technical)

National Electronics Complex of Pakistan (NECOP)

Building engineering automation tools and AI systems, including an on-premises RAG assistant and GENYSYS. Work spans Python, C/C++, FastAPI, HDL tooling, and system integration.

Jan 2022 — Present

Software Developer

Trivor Software

Desktop application development in C++ and .NET, together with .NET-based API development.

Jul 2021 — Jan 2022

Software Developer

Avanza Solutions

Developed desktop and middleware software for e-banking applications.

Jan 2021 — Jun 2021

Embedded System Engineer

STech.ai

Worked with NVIDIA Jetson systems for AI applications, particularly communication between web applications and edge devices.

Nov 2020 — Sep 2021

04 / EDUCATION

Where it started. Where it's going.

Strong computer engineering foundations, with a current focus on trustworthy and reliable AI systems.

MSFeb 2025 — Expected Apr 2027

MS in Computer Science

COMSATS University Islamabad · Wah Campus

Researching retrieval, persistent memory, and evaluation for LLM-based agents through EA-MEM.

3.91 / 4.00CGPA

BESep 2015 — Jul 2019

BE in Computer Engineering

National University of Sciences and Technology (NUST) · CEME

Final-year project combining CNN design with FPGA implementation for sign language recognition.

Rector's Gold MedalFinal-year project

05 / ABOUT ME

I care about systems that work outside the demo.

My path into AI research began with computer engineering and software development. Building real systems taught me that a good model is only part of the problem: reliable retrieval, changing context, and careful evaluation matter just as much.

Today, I study those questions through EA-MEM and other projects. I'm especially interested in the reliability of long-horizon agents, persistent memory, and retrieval-augmented systems.

Research areas

LLM agents · RAG · Information retrieval · Long-term memory · Agent evaluation

Technical toolkit

Python · C/C++ · FastAPI · ChromaDB · SentenceTransformers · Verilog · Vivado · Embedded systems

06 / GET IN TOUCH

Let's work on questions worth solving.

I'm open to PhD research opportunities and conversations about LLM agents, memory, retrieval, and dependable AI systems.

Send me an email