Saim Ahmad

Full Stack Software Engineer

• StudyScavenger — Engineered a HIPAA-aligned clinical research portal; built an AI eligibility engine (OpenAI + RAG) that reduced manual patient-to-trial matching by 70% and improved cohort accuracy by 45%. Deployed on AWS ECS with auto-scaling, achieving 99.9% uptime.
• LawAI — Built a full-stack legal intelligence platform, architecting a PostgreSQL vector store indexing 10,000+ judgments; integrated LangChain + RAG, slashing legal research time by 60% and cutting manual case review by 50%.
• DogWalker — Designed a 3-pass AI scheduling algorithm that reduced staff assignment conflicts by 55%; optimised routes via Google Maps API, cutting average travel time per shift by 30%.
• Ecommerce Store — Architected and delivered a containerized Django REST + React platform; established full CI/CD via GitHub Actions, reducing deployment time by 65% and achieving 100% zero-downtime releases.

Associate Software Engineer

• Meeting & Chat App — Architected real-time messaging via Redis Channels + WebSockets, achieving sub-50ms delivery; provided an AI assistant (OpenAI API) that eliminated manual summaries, cutting follow-up time by 40%.
• NearYou — Built an AI compatibility scoring engine (OpenAI API) that boosted relevant buddy-match rate by 35%; added real-time location features, increasing platform engagement by 28%.
• Email Intelligence Platform — Developed an enterprise enrichment system integrating 5+ third-party APIs; enforced OpenAI-based deduplication, improving data quality by 50% and reducing manual cleanup by 75%.
• Automation Scripts — Constructed 4 automation scripts (OpenPhone, auto-trading, NBA pipeline, furniture catalog) using Selenium + Playwright, eliminating manual data-collection effort by 80%.

Python/Django Developer Intern

• Multi-Brand Marketplace — Shipped and Dockerized a marketplace with independent vendor storefronts, reducing deployment overhead by 40% and achieving zero-downtime across all updates.
• Blood & Organ Donation Portal — Built a real-time donor/recipient matching portal with automated alerts, accelerating donor-recipient response time by 60%.
• I/O Optimization — Optimised I/O-bound operations with Python threading, boosting processing throughput by 45%.

Python Developer Intern

• Auto Scrape — Devised a BeautifulSoup scraping pipeline populating database models in real time, eliminating manual data entry by 90%.
• Smart CV Generator — Iterated a form-driven PDF CV generator (PDFKit + Pillow), slashing document production time from hours to under 2 minutes.

Adaptive Retrieval-Augmented Generation (ARAG)

Full-stack AI assistant built under the guidance of Dr. Mohamed Ahmed Sherif, Senior Postdoctoral Researcher:

• Designed and iteratively improving a full-stack AI assistant that lets users upload documents, create knowledge bases, and ask questions with streamed, citation-backed answers over their private data.
• Implemented and currently optimizing an adaptive RAG pipeline that classifies query intent and dynamically chooses retrieval, summarization, reranking, and generation strategies instead of relying on a static retrieval workflow.
• Combined and continuously refining dense vector search, TF-IDF/BERT retrieval, adaptive chunking, ChromaDB indexing, FastAPI APIs, React UI, SQLite persistence, and local/cloud LLM generation through Ollama and Gemini.

EduAI — Adaptive Learning Platform

GPT-4 powered adaptive learning platform built under Prof. Christian Koldewey:

• Conversational AI Interface — Authored a GPT-4 powered conversational AI interface, cutting student question-resolution time by 55% vs. traditional forum support.
• Adaptive Learning Engine — Implemented an adaptive engine that ingests PDF course material and auto-generates personalised remedial lessons (math formulations, flowcharts, course recommendations), reducing knowledge-gap closure time by 40%.

MSc Computer Science

BSc Computer Science

GPA: 3.55 (1.68 in German grading system)