Senior Data Scientist

Senior Data Scientist

6+ months contract

Yes, 100% remote

MS Dept of Child Protection Services

Position Description and Job Skill Set

The Mississippi Department of Child Protective Services seeks a senior data scientist to support a proof-of-concept demonstration using natural language processing and other machine learning methods to improve the intake process. This work is critical to demonstrate the potential of the latest technology to improve the lives of children at risk. The Senior Data Scientist will be responsible for overseeing and supporting the development, implementation, and testing of statistical models, integration of NLP, and refinement and testing of the prototype. In addition, algorithmic trade-offs will be evaluated, and guidance provided to ensure the State’s objectives are satisfied. The Senior Data Scientist will work closely with State stakeholders and technical team members to ensure the quality of the results and that the derived methods are transparent, statistically sound, relevant, and documented.

Key Responsibilities

• Create a Development Framework

o Establish a framework for the execution of technical tasks within the proof-of-concept. The framework will consist of task breakouts, milestones, and deliverables

o Identify critical milestones related to information, receipt of data, testing, and delivery.

o Identify key risk factors and means of mitigation.

• Current Processes & Technology

o Participate in critical discussions involving current intake workflows, how decisions are made based on information from the intake process, and the allocation of State labor

o Contribute to the identification of intake process shortcomings and prioritize based on their impact on children and State resources.

• Devise New Intake Approach Using New Technologies
• Bachelor’s, Master’s, or Ph.D. in computer science, mathematics, engineering, physics, or related field.
• Have participated in US Federal Gov’t data science programs requiring TS/SCI clearance, delivering solutions requiring the combination of geospatial disciplines and pattern of life, and Social network connections.
• Data engineering expertise, with demonstrable experience custom building programs processing in excess of 700 Million records in less than :30min, on a highly frequent, reoccurring basis.
• Proven expertise working with CCWIS data attributes to predict child welfare outcomes, including but not limited data attribute selection, data clean up and statistical tuning.
• Extensive knowledge of statistical algorithms, machine learning, and adaptive systems.
• Prior history of designing and building machine learning algorithms from the ground up.
• Experience with making technical trade-offs between algorithmic approaches. based on collective errors, computational time, scalability, and outcomes.
• Prior success in developing optimal non-rule-based decision-making systems where the inputs are stochastic.
• Successful history of converting social processes and human decision-making into computational models that yield improved results.

o Lead the development of a new intake process that leverages natural language processing and other machine learning algorithms.

o Identify the functional blocks and reconcile their contributions to solving the prioritized shortcomings.

o Evaluate architectural and computational implementation trade-offs for each functional block. The evaluation should consider risk from the standpoints of technical, schedule, and security.

o Evaluate trade-offs of using different data sources, including existing systems, sample data, simulated data, or other alternatives.

o Document the final approach for transparency.

• Design Review(s)

o Create the framework for the design review process.

o Lead the design review and evaluate

 The functional design with respect to resolving prioritized intake process shortcomings, and the impact on children and State resources.

 Technical, schedule, data security, and other risk factors.

 Source of data and its usefulness in demonstrating the efficacy of the approach.

 Proposed methods of test and demonstration.

o Documentation of the process for transparency.

• Implementation of Proof-of-Concept

o Oversee the implementation of the prototype by conducting weekly status updates and, when appropriate, gate reviews.

o Provide guidance when needed to mitigate risk and remove technical or administrative roadblocks.

• Conference Room Demonstration

o During the course of 3-4 days, provide conference room support to demonstrate that shows how the prototype application can improve child outcomes and reduce State resources.

o Capture key stakeholder comments regarding technical aspects of the application.

• Roadmap

o Contribute to the development of a roadmap that illustrates how the developed technology could be integrated into the State’s ecosystem of technologies and processes.

• Agile Development Process

o Contribute to the Agile development process to ensure the success of the project.

Required Skills/Experience:

• Bachelor’s, Master’s, or Ph.D. in computer science, mathematics, engineering, physics, or related field.
• Have participated in US Federal Gov’t data science programs requiring TS/SCI clearance, delivering solutions requiring the combination of geospatial disciplines and pattern of life, and Social network connections.
• Data engineering expertise, with demonstrable experience custom building programs processing in excess of 700 Million records in less than :30min, on a highly frequent, reoccurring basis.
• Proven expertise working with CCWIS data attributes to predict child welfare outcomes, including but not limited data attribute selection, data clean up and statistical tuning.
• Extensive knowledge of statistical algorithms, machine learning, and adaptive systems.
• Prior history of designing and building machine learning algorithms from the ground up.
• Experience with making technical trade-offs between algorithmic approaches. based on collective errors, computational time, scalability, and outcomes.
• Prior success in developing optimal non-rule-based decision-making systems where the inputs are stochastic.
• Successful history of converting social processes and human decision-making into computational models that yield improved results.