My Work
Professional experience, research, and achievements
Experience
Intelligent Operations Engineer
- →Engineered Bash-based monitoring pipelines for distributed systems across 50+ servers, reducing downtime by 70%+ and improving system reliability to 99.95% uptime
- →Architected real-time fiber fault detection system using machine learning, reducing detection latency from hours to under 30 minutes, enabling faster incident response and customer satisfaction improvements
- →Administered role-based access control and secret management for 500+ users across distributed systems, ensuring compliance with security policies while maintaining operational efficiency
AI Software Engineer
- →Built end-to-end investment tagging platform using Next.js and FastAPI that automated Affinity CRM workflows and reduced manual data entry by 90%, enabling rapid portfolio company screening and analysis at scale
- →Engineered a production-grade RAG system using Qdrant Vector Database and OpenAI embeddings for autonomous company analysis, enabling semantic search across 10,000+ funding documents
- →Deployed fully containerized microservices on Render using Docker and implemented CI/CD pipelines, reducing deployment time from 2 hours to 15 minutes
Engineering Fellow — Impact Projects Program
- →Built and deployed AI-powered research systems using Qdrant, GPT-4, and vector search to classify 8,000+ construction tech companies with 85%+ accuracy, creating the foundation for impact investment decisions
- →Developed full-stack internal platform using Next.js and FastAPI with advanced filtering and tagging features that streamlined startup analysis, reducing research time by 70% and enabling 5x faster deal flow screening
- →Collaborated with international teams across 6 countries on sustainable development and affordable housing technology initiatives aligned with UN SDGs, directly supporting 200+ social impact startups
Research Intern
- →Researched prompt optimization techniques for hallucination reduction in LLMs, analyzing 500+ prompt variations and achieving 35% improvement in factuality scores on benchmark datasets
- →Evaluated adversarial prompting strategies to improve model robustness against jailbreak attempts, contributing to EleutherAI's safety research initiatives
Machine Learning Researcher
- →Designed a multilingual red-teaming framework in collaboration with Stanford NLP Group and Georgia Tech, covering cultural biases and contextual failures in LLMs across diverse geographies
- →Curated and validated a comprehensive dataset of 160,000+ prompts across 8 languages and 200+ cultural domains, enabling research on cultural representation in AI systems
- →Co-authored peer-reviewed research paper on culturally contextual red teaming for LLM alignment, contributing to the broader conversation on equitable AI development
Research Fellow
MIT Computer Science & Artificial Intelligence Laboratory (CSAIL)·Boston, MA·July 2024 – November 2024
- →Trained a CNN via PyTorch for 3D cortical mesh segmentation achieving 70% dice coefficient accuracy on benchmark neuroimaging datasets, advancing medical imaging analysis
- →Built comprehensive evaluation benchmarks for deforming 3D surfaces measuring mesh quality through geometric signals including Hausdorff distance and surface smoothness metrics
- →Co-authored two peer-reviewed publications on cortical segmentation and geometric deep learning presented at MIT Summer Geometry Initiative 2024
AI Accelerator Program Reviewer
- →Evaluated 80+ early-stage African AI startups on technical merit, business viability, and innovation potential, providing detailed feedback on technical architecture and go-to-market strategies
- →Supported equitable selection of Meta-backed AI teams focused on generative AI and LLMs, ensuring diverse representation and supporting 15 cornerstone startups through the accelerator program
Awards & Honors
Presidential Grant Award for Academic Excellence
January 2025