Chief Software Architect Resume Philadelphia, PA - Hire IT People

SUMMARY:

Data science expert in machine learning, pattern recognition, and design of distributed systems.
Exceptional command of mathematical concepts employed by machine learning methods, coupled with excellent understanding of algorithms design and their complexity.
Effective project manager. Proficient software developer and systems architect.
Excellent written and oral communication skills.
Seven years of Confidential and development experience in the industry.
Spent five years as principal data scientist at data analytics startup: developed machine learning models in Python for time - series tagging and segmentation; implemented appropriate visualization dashboards for these models using Pandas, Numpy, Jupyter and Bokeh libraries.
Doctoral Confidential focused on large-scale information and content retrieval; and involved learning of sequence embedding in feed-forward deep neural networks using Torch library.
Developed sensor fusion and forecasting models in Java for travel delays on highways and arterial roads.

RELEVANT TECHNOLOGIES:

PROFESSIONAL EXPERIENCE:

Chief Software Architect

Confidential, Philadelphia, PA

Responsibilities:

Formulated, implemented and optimized principal data analysis model for DRA: online-trained model for segmenting time-series into “normal” and “atypical” categories. Atypical category indicate observations that are considered anomalous when compared to observations from training data.
Developed robust detector of monotonous time intervals of arbitrary size in time-series. Implemented in Python using Numpy and Pandas libraries. Method relied on singular spectrum analysis and automatic scale selection in a scale-space pyramid.
Designed an “if-then” framework based on evaluating arbitrary algebraic expressions of time-series (i.e., process) variables. The framework facilitated use of user-defined expressions to encode logic of an “if-then” rule. Multiple rules can then be organized to form a decision tree.
Evaluation of these expressions was implemented using thin wrappers around pandas.Series and pandas.DataFrame
Visualization prototype was implemented as a Bokeh application
Re-formulated principal analysis model as a convolution operator. This allowed its seamless integration with the “if-then” framework, where segmenting a time-series is equivalent to solving a linear system.

Data Engineer

Confidential

Responsibilities:

Designed data model for segmented time-series for fast rendering in a client’s browser:
Supported rendering of time-series with millions of points at arbitrary date ranges
Wrote Python code for saving segmented time-series as level-of-detail pyramids
Wrote core rendering logic in D3.js
Designed and implemented DRA backend as fault-tolerant daily task processor
Dask library was used for task scheduling and execution
Designed Python-based packager and installer tool to automate deployment and maintenance of DRA

Project Manager

Confidential

Responsibilities:

Developed feature development protocol that formalized the process of building DRA features by Dev. The procedures covered the progress of each feature from its inception in UX/Product group to its deployment and support by the Dev group
Assigned work to Dev according to feature priority and outstanding client requests
Estimated release dates for new features
Tracked progress of outstanding client requests to meet promised timeframes
Tracked and evaluated performance of the Dev team

Client Support Engineer

Confidential

Responsibilities:

Implemented over a dozen of vendor-specific adaptors for DRA to load data from time-series databases (a.k.a., “historians”) for process control data
Worked with clients to create DRA deployment and support plan that would meet their operational protocols for network and computing resources
Travelled onsite to assist high-value clients with deploying DRA to their networks

Senior Software Developer

Confidential, Wayne, PA

Responsibilities:

Was responsible for Confidential, development and implementation of algorithms to be utilized by the real-time traffic monitoring system.
Developed short-term future traffic flow prediction system based on finding similar patterns in current and historical traffic flows.
Developed system for traffic flow speed and travel time estimation that fused together information on traffic conditions from various sources: stationary microwave sensors, toll-tag readers, actual moving vehicles with Confidential receivers, incident reports, and historical speed averages.
Both systems have undergone nationwide evaluations for accuracy via numerous test drives in major metropolitan areas.

Graduate Assistant

Confidential, Phila, PA

Responsibilities:

Conducted Confidential in pattern recognition and information retrieval under supervision of Confidential .
Collaboration with researchers from Confidential Labs has resulted in development of novel model for exploiting spatial evidence among sequences of features in supervised latent spaces, which has become major contribution of my dissertation work.
Having completed M.S. degree before start of Ph.D. program enabled to commit first three years of the program to study advanced topics in Mathematics.
During these years participated in development phases of web portal sprucecommunity.org, funded by Confidential Laboratory ( Confidential ), for collaborative engineering of software intensive systems.
Confidential team developed expert discovery system that facilitated keyword-based semantic retrieval of researcher profiles ranked by their expertise in query topic.

Assistant (Intern)

Confidential, Princeton, NJ

Responsibilities:

Conducted Confidential in large-scale text and image classification under supervision
Developed system for large-scale opinion mining and text categorization.
Distributed training framework for large-scale classification was implemented in C and Lua using Torch machine learning library.
MySQL database was used for synchronization and message passing.