|
|
About Qtek
|
|
|
|
|
|
|
|
Areas Of Expertise:
Skills: 2D and 3D Computer Graphics (BSP trees, collision detection, 3D graphics engine design), A.I. (Neural Networks, Reinforcement Learning Algorithms, Genetic Algorithms), OOP, Inverse Kinematics for robot controllers, Networking of VR environments.
Languages: C/C++, Microsoft Visual C++, Borland C++ Builder, Java, OpenGL, DirectX, MFC, Microsoft Visual Basic.
|
|
Resume:
Summary: Highly qualified individual with exceptional technical skills and research capability.
Qualifications: Master&'s degree in Computer Science with Distinction
Background to University Experience and Qualifications:
Master&'s Degree in Computer Science (1st Class Pass)
My master&'s degree was undertaken purely as a research degree and, thus, was evaluated entirely on the following project:
Design, Evaluation and Comparison of Evolution and Reinforcement Learning Models
Reinforcement Learning and evolutionary or genetic algorithms form what can be considered to be the leading edge of modern A.I. For this reason I undertook the above research in order to investigate the adaptive learning nature of these models. An additional aspect of the research was the goal of extracting the adaptive advantages of reinforcement learning and evolutionary based models in order to produce hybrid models capable of a superior adaptive nature.
In order to achieve these goals a simulation environment and research tool was created. This was done using c++ to provide an object orientated framework incorporating neural networks, reinforcement learning and evolutionary or genetic algorithm models. In addition, data extraction tools were also created in order to monitor the performance of the models. The extracted data for each experiment being processed and displayed by these tools after being averaged over 30 trials in order to produce statistically relevant data with acceptable standard deviations.
These experiments took place in a simulation context or environment was constructed in order to provide a task that was to be achieved by the previously mentioned models. The task itself was to be achieved by agents populating the virtual environment, each agent consisting of a neural network, reinforcement learning and/or an evolution based learning model.
The results of my research demonstrated that by combining the adaptive advantages of reinforcement learning and evolutionary based models, in combined hybrid models, learning of a superior adaptive nature can be achieved. In addition, additional experiments were performed taking this work further by utilising the adaptive advantages of two forms of evolution (Darwinian and Lamarckian evolution) in order to produce a remarkably superior hybrid model, capable of optimising the nature of learning required according to the conditions of the environment.
Honours Degree in Computer Science
Honours Degree Project:
Multi-user, Interactive Virtual Reality via the Internet.
This project was undertaken in order to investigate the issues involved in the distribution of large-scale, widely-distributed Virtual Reality environments over networks, like the internet.
Such issues included:
· Minimising latency and its effects.
· Maintaining the shared worlds consistency.
· Maximising the number of users that can exist and interact in the shared world.
· Shared resource contention resolution.
· Collision detection inconsistencies caused by the effects of latency.
Solutions were proposed to the above issues and notable results achieved with the implementation of a dead-reckoning method, whereby the movements of objects are predicted rather than transmitted. This was shown to dramatically reduce the effects of latency, thereby noticeably improving the fluidity of the VR simulation.
Honours Graphics Project: Created a BSP compiler to optimally organise polygons into a data structure and viewer to read the BSP data structure and render the polygons in the appropriate order.
In addition to my honours project work, courses were undertaken in:
· Computer Graphics
· Parallel and Distributed Processing
· Multimedia
· Computer Hardware Interfacing
· Java
· Computer Networks
Undergraduate 3D Graphics Engine Project:
As an undergraduate I took a personal interest in the development of a 3D graphics engine. This culminated in the design of a graphics engine providing the following useful features:
· 3D animation objects.
· Dynamic creation and destruction of objects in the environment
· Collision detection
· Camera or viewpoints that can be attached to objects and move relative to them.
· Objects themselves can be defined relative to other objects.
Contract Work:
Pipe Bending Tube Visualisation Project:
This project involved the creation of a tube visualisation application for a client in the pipe bending industry. The application allows for the visualisation and manipulation of tube structures as created by pipe bending machines. C++, MFC and OpenGL were used create a DLL that could be utilised by a Visual basic Application to view and manipulate the tube structures.
The following features were provided:
· Rotation, scaling and movement of the tube structures.
· Selection of any particular segment as the tube&'s centre point of rotation.
· Ability to create circular, oval, rectangular and user definable cross-sections.
· Variables defining the tube&'s appearance can be specified according to the requirements of the user. These include: the tube&'s radius, level of detail (number of vertices along the tube&'s circumference and number of rotations used to produce a desired bend), length of top and bottom flat sections for oval tubes with flat sides.
· Ability to automatically scale the tubes according to the size of the viewing window, regardless of the tube&'s dimensions.
Robotic Arm Controller Project:
Developed inverse kinematic calculations to control a 6 DOF robotic arm and applied these to a virtual robot simulation for the previously mentioned client who is interested in the use of robotic arms as part of the pipe bending process.
This project involved the following:
· Inverse kinematic calculations were developed in order to provide the correct configuration for a 6 DOF robotic arm.
· Created a virtual robot in order to test the inverse kinematic calculations and movement of the arm.
· In addition, the inverse kinematic calculations were designed in order to ensure that the robot&'s end point is maintained at a particular, specified orientation. For example, it may be desired that a spray painting tool that is to be held by the robot is maintained at a level orientation with respect to the ground.
· Applied the inverse kinematic calculations and virtual robot to a simple task, that of picking up and moving randomly scattered boxes to a particular position.
Other Interests: Gym, touch-rugby, Part-time PhD investigating the possible use and popularisation of
physics hardware cards and the development of standards, like the OpenGL standard, in order
to program such cards.
|
|
Top Coder Competition Raw Score
|
| Average Job Size |
|
Average Coder Rating |
|
Number of Completed Jobs |
|
Penalty for Missed Status Reports Warnings * |
| $283.33 |
x |
10 |
x |
3 |
- |
(0 warnings x 1000 penalty) = 0 |
Raw Score Total: 8,500
* Penalty rule enacted on 10/8/2001 and applies only to warnings sent after that date.
Note: all top coder competition stats are only updated once a day between midnight and 1:00 EDT.
|
|
|
|
|
|
|
Rent A Coder Work History
|
All ratings and comments are the opinion of the poster and do not reflect the official view of Rent A Coder.
|
Rated For |
Rated By |
Rating |
Bid Amount |
Rated On |
| |
Tube_3 |
MCW |
10
(Excellent)
|
$400.00 |
February 19, 2002 2:46:58 PM EDT |
| |
The rating and a repeat customer says it all!!! |
| |
| |
Tube_2 |
MCW |
None Given
|
$200.00 |
December 12, 2001 6:42:12 AM EDT |
| |
|
| |
| |
Tube |
MCW |
10
(Excellent)
|
$250.00 |
December 6, 2001 11:30:32 AM EDT |
| |
Absolutely first class work and a great person to deal with. |
| |
| |
------ |
| Average: |
|
10.00 |
|
|
|
|
Rent A Coder upholds the rigorous business practices required to be both a BBB member and Square Trade vendor.
- All customer issues addressed within 2 days
- Openly disclosed pricing and return policies
- Participation in mediation at buyer request
- Superior selling track record
This site is verified through its parent company, Exhedra Solutions, Inc.
|
|
|
