Competencies

• Grant Proposal Writing
• LLMs
• Knowledge Graphs
• Neuro-symbolic AI
• Retrieval Augmented Generation
• Full Stack Web Development
• GPT-4

Achievements

• Principal Investigator on NSF SBIR research initiative - conceptualized novel technical approach and authored successful Project Pitch and subsequent Phase I research proposal.
• Conducting comprehensive literature reviews of state-of-the-art hallucination mitigation approaches, and developing novel integration strategies using advanced retrieval and knowledge representation techniques.
• Leading the development of Govskills.io, leveraging AI to streamline complex job search processes for federal positions.

Competencies

• LLMs
• Llama-3
• Supervised Finetuning
• Evaluation
• Research Methodology
• Lamini
• Automated Unit Test Generation

Achievements

• Conducted comprehensive literature review and led research initiative exploring advanced LLM fine-tuning approaches for automated unit test generation.
• Performed systematic evaluation of leading open-source models for code generation, benchmarking their performance on complex C testing scenarios.
• Engineered automated feedback mechanisms that achieved 30% improvement in test pass rates through systematic iteration and optimization.

Competencies

• LLMs
• GPT-4
• Retrieval Augmented Generation
• RAGAS
• Langchain

Achievements

• Led research and development of Federal Regulation Explorer, conducting literature review of state-of-the-art RAG techniques and implementing them for regulatory text analysis.
• Iteratively enhanced baseline RAG system through systematic experimentation with advanced retrieval strategies (re-ranking, multi-stage retrieval, and fusion techniques), achieving 20% improvement in RAGAS metrics.
• Designed and implemented systematic evaluation methodology incorporating multiple RAGAS metrics (context relevance, retrieval precision, faithfulness), enabling quantitative analysis of each architectural enhancement’s impact.

Competencies

• Data-Driven Decision Making
• Automation & Efficiency
• Cloud Architecture & Scalability
• System Reliability & Incident Management
• Leadership

Achievements

• Doubled platform throughput by designing a high-availability AWS infrastructure, mirroring the demands of deploying scalable machine learning models.
• Reduced incident response times by 60% by implementing a comprehensive monitoring solution (ELK stack, Cloudwatch) for a 100+ node distributed system, critical for the continuous operation of ML applications.
• Reduced deployment cycles by 50% (-25min) developing an automation framework with Ansible & Terraform, streamlining configurations and deployments, essential for agile development and machine learning workflows.
• Enabled dynamic scaling of core services by spearheading the transition to containerized applications using Docker and AWS ECS, crucial for agile machine learning model development and testing.
• 0-1 design and implementation of a scalable microservices architecture, demonstrating an understanding of scalability essential for data-intensive machine learning operations.
• Reduced search latency from 6 secs to under 1 sec by strategic caching and switching from Postgres to Elasticsearch, optimizing query performance and user experience.

Competencies

• Performance Optimization
• API & Data Integration
• AWS Cloud Engineering

Achievements

• Redesigned the system to a microservices-based architecture, reducing response time by 30%.
• Integrated Stripe, Twilio, and SendGrid APIs to enhance functionality and UX.
• Deployed cloud-based web apps on AWS, ensuring scalability and reliability.

Competencies

• Problem Solving
• Continuous & Adaptive Learning
• Collaboration & Teamwork

Achievements

• Developed RESTful APIs for the platform's mobile app backend, resulting in improved app functionality and user experience.
• Revamped the legacy pricing module, making it more extensible to newer pricing models.
• Performed routine bug fixes and housekeeping tasks to ensure the smooth operation of the platform.