Tanweer Ashif

About

Data science/AI intern at University at Buffalo Business Services who built and deployed production systems spanning classic ML and LLM assistants. Delivered real-time competitor intelligence for a Cornell-partnered, $1B beverage launch by scraping/cleaning 5,000+ SKUs and deploying models via API, then built a domain-aware LLM assistant to modernize Excel-based workflows with strong grounding, privacy controls, and sub-5s latency.

Experience

Education

Key Strengths

• Built production ML system for real-time competitor intelligence from messy e-commerce data (5,000+ SKUs)
• End-to-end data pipeline development: scraping, deduping, normalization, feature engineering, and deployment via API
• Trained and compared multiple ML models (9+ including Random Forest and XGBoost)
• Deployed scalable services using Docker + Kubernetes (stateless FastAPI services, scheduled refresh/evaluation jobs)
• MLOps discipline with MLflow experiment tracking and controlled model promotion via registry
• LLM domain-grounding via task-specific dataset curation and fine-tuning; tuned temperature/top-p/top-k for factuality vs coherence
• Handled privacy constraints for internal records with data scrubbing and restricted-access pipelines
• Performance engineering for LLM workflows using deterministic templates, batching, and caching to keep latency under ~5 seconds
• Strong stakeholder collaboration: translated non-technical marketing needs into metrics, dashboards, and usable daily tools

Reference Highlights

• took operation from nothing to a very robust product
• calming presence
• tolerant and easy to work with
• kind and genuinely nice to work with
• well liked by coworkers
• well respected
• patient when explaining technical topics
• effective communicator with non-technical stakeholders
• Excellent at translating non-technical requirements into working solutions
• Strong collaborator; team worked seamlessly
• Highly supportive; continues to consult and help new interns
• Clear communicator of technical concepts to non-technical stakeholders
• Went above and beyond
• Reliable and deadline-driven (hit every deadline)
• Effective under pressure; handled day-of issues successfully
• Built a project designed to live on and be extended