Muhammad.

Muhammad Aarash Abro

ML Engineer | Scaling RAG Pipelines & Automating Enterprise Workflows

Overview

Machine Learning Engineer specializing in high-scale AI systems, from processing 20+ TB of financial data to building agentic workflows for global firms like BDO and Warner Bros. Expert in RAG, RL, and time-series modeling with a focus on measurable efficiency gains and research-backed solutions.

Experience

2023 – Present

WORK

Research Lead (ML for Chemical Reaction Modeling)

Auxilart (Industry-Academic Collaboration)

Leading research on organic reaction mechanism identification under real-world data heterogeneity.

•Organic reaction mechanism identification under real-world data heterogeneity (Under Review, ESCAPE 36): Built Transformer-based encoder for irregular chemical trajectories and sparse-autoencoder based domain adaptation module; achieved 99.8% accuracy (20% masking) and 93.4% (40% masking); implemented full ODE-based data-generation pipeline.

Key Achievements

→Achieved 99.8% accuracy (20% masking) and 93.4% (40% masking) in reaction mechanism identification.

TransformersSparse-autoencodersDomain AdaptationODE-based data-generation

Jul 2022 – Present

CONTRACT

Machine Learning Consultant

Zeta Solutions

Worked on multiple high-impact AI automation projects across various industries including legal, finance, and entertainment.

•ATR SmartProcedures — AI-Driven Factory Documentation: Automated digitization of factory equipment manuals into proprietary formats using LLMs, enabling seamless integration with internal software, reducing turnaround time per-table by 80%.

•UHY Prime HK — Audit Automation: Automated auditing by leveraging LLMs and VLMs to match invoices and documents with transactions, verify accuracy, and flag discrepancies, reducing number of man-hours required by up to 90%.

•Warner Bros. Discovery — Post Production Video Assistant Automation: Built computer vision models for automated text detection in videos, streamlining translation and post-production in the entertainment industry.

•Whichdraft — AI Legal Assistant for Contract Generation: Developed agentic workflows for AI-generated wizards to assist novice lawyers in drafting standard contracts to reduce time spent on boilerplate work.

•Enviro AI — Environmental Compliance Chatbot: Built an agentic reasoning system using TCEQ knowledge bases to automate environmental compliance reviews and permit applications for clients such as ExxonMobil and Dow Chemicals.

•BDO Global — Efficiency Modeling and Routing Optimization: Extracted oil-pump efficiency curves from legacy documents using computer vision and curve fitting, facilitating optimized oil flow routing.

Key Achievements

→Reduced turnaround time per-table by 80% for factory documentation digitization.

→Reduced man-hours required for auditing by up to 90% through LLM/VLM automation.

LLMsVLMsComputer VisionAgentic WorkflowsCurve Fitting

Sep 2024 – Aug 2025

WORK

Research Assistant

Intelligent Machines & Sociotechnical Systems Lab, LUMS

Lahore, Pakistan

Supervised by Dr. Hassan Jaleel. Focused on financial decision making and regime shifts.

•No-Regret Portfolio Optimization (Accepted, AI4DF Workshop @ ICAIF 2025): Developed a multi-agent no-regret learning framework for financial decision making and regime shifts; outperformed S&P 500 and gold on risk-adjusted returns; implemented large-scale simulation and stress-testing pipeline.

Key Achievements

→Outperformed S&P 500 and gold on risk-adjusted returns using a multi-agent no-regret learning framework.

Multi-agent learningNo-regret learningSimulation pipelineStress-testing

Jun 2024 – Jul 2025

WORK

Research Assistant

Electrical Engineering Department, LUMS

Lahore, Pakistan

Supervised by Dr. Muhammad Tahir. Focused on financial modeling and domain adaptation.

•Self-Supervised Financial Modeling with RL: Developed multiresolution time-series representation model using transformer feature generator and PPO-guided pretext tasks; processed 20+ TB of data with Python–C++ pipelines on Cloud Run, BigQuery, and Cloud Storage.

•Unsupervised Domain Adaptation with Sparse Autoencoders: Designed SAE-based domain-invariant encoder improving cross-domain and few-shot performance for vision backbones.

Key Achievements

→Processed 20+ TB of data using Python–C++ pipelines on cloud infrastructure.

PPOTransformersPythonC++Cloud RunBigQueryCloud StorageSparse Autoencoders

Project Portfolio

PERSONAL

Present

Scalable RAG Pipeline for Secure Knowledge Processing

Developer

Developed a customizable Retrieval-Augmented Generation (RAG) pipeline to process dense embeddings at scale, enabling secure and cost-efficient knowledge retrieval for businesses. Reached embedding throughput of 10 Million Tokens Per Second throughput with $30 compute budget by optimizing inference through maximal threading of I/O bound processes and parallel serverless processing for compute-bound processes.

RAGEmbeddingsThreadingServerless Processing

ACADEMIC

Present

Job Posting Analytics

Developer

Designed a data pipeline to scrape job postings and extract key skill requirements using Small Language Models (SLMs). Built a job board to aggregate frequently demanded skills by position, simplifying career progression and skill acquisition pathways.

Small Language Models (SLMs)Web ScrapingData Pipeline

Education

Sep 2021 – Jul 2025

B.S. Computer Science

Lahore University of Management Sciences (LUMS)Specialization in Computer Science

GPA: 3.48

Relevant Coursework

Reinforcement LearningAdvanced Topics in MLMultiagent SystemsRoboticsDistributed SystemsData ScienceMachine LearningDeep learning

Impact & Recognition

Patents & Publications

2025 · PUBLICATION

No-Regret Portfolio Optimization

Accepted at AI4DF Workshop @ ICAIF 2025.

Team:Dr. Hassan Jaleel

PUBLICATION

Organic reaction mechanism identification under real-world data heterogeneity

Under Review, ESCAPE 36.

Skills & Interests

Python

TECHNICAL

Large-Scale Time-Series Modeling

TECHNICAL

PyTorch

TECHNICAL

Hugging Face Transformers

TECHNICAL

NumPy

TECHNICAL

Pandas

TECHNICAL

Large Language Models

TECHNICAL

Retrieval-Augmented Generation

TECHNICAL

Reinforcement Learning

TECHNICAL

scikit-learn

TECHNICAL

C/C++

TECHNICAL

No-Regret Learning

TECHNICAL

BigQuery

TOOL

Cloud Storage

TOOL

Cloud Run

TOOL

SQL

TECHNICAL

Self-Supervised Representation Learning

TECHNICAL

TensorFlow

TECHNICAL

Domain Adaptation

TECHNICAL

JavaScript/TypeScript

TECHNICAL

Bash/Shell

TECHNICAL

Docker

TOOL

PostgreSQL

TOOL

Redis

TOOL

Go

TECHNICAL

Let's Connect

I'd love to hear from you. Feel free to reach out.

These unlock once you finish onboarding and publish your profile.

Talk to AI Muhammad

Online

I am an AI representation of Muhammad. Ask me anything about their experience, skills, or working style.