Summary
Overview
Work History
Education
Skills
Websites
Leadership Interests
Timeline
Generic

Emmanuel C. Shedu

Oakland,CA

Summary

Results-driven machine learning software engineer with advanced experience in large-scale model training, fine-tuning LLMs and diffusion models, and deploying full-stack AI solutions. Currently at Meta, improving core ML pipelines across Instagram and Facebook. Founder of MangaReadAI, engineered a state-of-the-art manga panel segmentation and anime-style transfer pipeline using Stable Diffusion, LoRA, and ControlNet. Proven track record in deep learning, model optimization, scalable infrastructure, and cross-functional delivery. Skilled in [Skill] with a documented history of discovering methods to intelligently use data to enhance user experience. Effectively researching techniques for novel approaches to problems, developing prototypes to assess the viability of the approach, and deploying applications into production yielding insights to expand customer-consciousness.

Overview

11
11
years of professional experience

Work History

Machine Learning Engineer - Trust, Integrity & Safety

Meta Platforms
05.2024 - Current
  • Led age-prediction ML model development deployed across Facebook and Instagram.
  • Increased ground-truth quality by 35% using novel FB status update label sourcing.
  • Reduced pipeline compute by 14% through targeted query optimization.
  • Engineering Champion: Led compute and manual effort reduction strategies.
  • Skills: PrestoSQL, PyTorch, Spark, Python, MLFlow

Founder & Lead Machine Learning Engineer

MangaReadAI
04.2024 - 06.2024
  • Company Overview: Solo Project
  • Built full-stack, AI-powered manga platform with panel segmentation, speech bubble detection, and style transfer for animation-ready assets.
  • Fine-tuned Stable Diffusion, LoRA, and ControlNet for high-fidelity anime/manga style transfer (e.g., Boruto, Solo Leveling), achieving robust generalization across frames.
  • Engineered training pipelines with smart cropping, data augmentation, edge detection, and color preservation; optimized for GPU/cloud compute efficiency.
  • Automated batch processing, checkpointing, logging, validation, and quality control at scale.
  • Integrated U-Net-based DeepPanel segmentation (outperforming OpenCV), and React-based comic reader frontend.
  • Resolved Keras 3/TensorFlow 2 compatibility issues and migrated legacy code to modern ML practices.
  • Solo Project
  • Key Accomplishments: Achieved high-style fidelity in both grayscale and color images for single-frame and vi2v inputs. Reduced training time while improving output through parameter tuning and targeted data curation.
  • Skills: Stable Diffusion, LoRA, ControlNet, PyTorch, TensorFlow, OpenCV, U-Net, GPU training, cloud ML, ML pipeline automation, React, Docker

Full-Stack Developer

IBM
07.2019 - 09.2023
  • Resolved 25–40 customer defects weekly by participating in daily triage calls with cross-functional teams, significantly reducing response time and boosting client satisfaction.

    Enhanced and maintained 15+ key product features, contributing to a 20% reduction in bug recurrence across production releases.

    Improved regression test coverage by 35% by designing, updating, and fixing automation scripts, leading to earlier defect detection and smoother CI/CD cycles.

    Reduced manual QA effort by ~40 hours/month through automation of repetitive deployment tasks.

    Led 5+ application deployments per quarter, ensuring seamless delivery of updates and new features with zero major rollbacks.

    Supported ongoing modernization of IBM Content Navigator, contributing to a 10% improvement in UI performance through frontend refinements and bug fixes.

    Skills: Java, JavaScript, HTML/CSS, Docker, Selenium, Azure

Research Assistant - DEPA Lab

University Lab
01.2014 - 12.2017
  • Project 1: Occupancy Tracking Using Visible Light Communication (VLC)
  • Engineered a visible light communication system using C++ on Arduino microcontrollers, enabling inter-device data transmission via LED infrastructure within indoor environments.
  • Developed modular hardware prototypes, including addressable LED transmitters and IR-based ID tags, for real-time room occupancy detection and grid mapping.
  • Increased system reliability by 25% through iterative testing, shielding, and circuit optimization for stable data transmission.
  • Mentored 5+ undergraduate research assistants, accelerating onboarding and lab productivity.
  • Project 2: Social Media Data Mining for Analytics & Security
  • Built a real-time data pipeline in Python to mine and store targeted Twitter data using both REST and Streaming APIs; persisted datasets into a MySQL database for sentiment and event analysis.
  • Automated database cleanup and optimization with a custom PHP cron-based system, maintaining data volume and structure for high-efficiency querying and downstream ML workflows.
  • Project 3: Autonomous Vehicle with Obstacle Avoidance & SLAM
  • Designed and assembled an autonomous vehicle prototype using lightweight chassis materials (cardboard, motors, and sensors) to minimize build cost while supporting embedded processing.
  • Overclocked a Raspberry Pi B to run real-time computer vision and RGB-D sensor processing for localization and collision detection.
  • Implemented a Kalman Filter-based SLAM algorithm for simultaneous mapping and localization in unknown environments.
  • Technologies: C++, Arduino, Raspberry Pi, Python, MySQL, PHP, OpenCV, LED-based communication, REST API, Kalman Filter, computer vision, embedded systems

Education

Master of Science - Engineering Management (CS focus)

John Hopkins
Baltimore, MD
12-2019

Bachelor of Science - Computer Engineering

Morgan State Univeristy
Baltimore, MD
12-2017

Skills

  • Stable Diffusion
  • Low-Rank Adaptation
  • ControlNet
  • PyTorch
  • TensorFlow
  • MLFlow
  • Data pipeline engineering
  • Cloud training (GPU)
  • Automation
  • Python
  • Java
  • C
  • JavaScript
  • PrestoSQL
  • PHP
  • Docker
  • Kubernetes
  • Jenkins
  • AWS
  • React
  • NodeJS
  • Selenium
  • CI/CD
  • Machine learning
  • Data mining

Leadership Interests

  • Social group leader & mentor
  • Digital artist (manga/comics)

Timeline

Machine Learning Engineer - Trust, Integrity & Safety

Meta Platforms
05.2024 - Current

Founder & Lead Machine Learning Engineer

MangaReadAI
04.2024 - 06.2024

Full-Stack Developer

IBM
07.2019 - 09.2023

Research Assistant - DEPA Lab

University Lab
01.2014 - 12.2017

Master of Science - Engineering Management (CS focus)

John Hopkins

Bachelor of Science - Computer Engineering

Morgan State Univeristy

Similar Profiles

CREATE PROFILE