Hikmet Bisen
Machine learning + mechanical engineering — Austin, TX, class of 2027
Selected work
In-ear multimodal sensing device
heart rate · respiration · voice · motionLinkRing — self-hosted health tracker
your data · your baseline · no subscription1983 BMW E30 restoration
chassis gone over seam by seam- FORGE — AI engineering analysis platform — RAG + multi-model routing, built solo
- Python banking system — 40/40 unit tests
- Claude usage dashboard — single-file Python, SSE live metrics
- Obsidian AI memory graph — personal knowledge architecture
- Roller coaster design (LESC)
- Code2College
- Architecture research (AIA)
Climbinglead 5.10d · top rope 5.11c · boulder V7 — lead-certified one month after first touching a wall (2026)
About
Machine learning + mechanical engineering, Austin, TX. Dual enrollment at Harmony School of Endeavor and Austin Community College — 4.56 GPA, class of 2027.
The work runs hardware to software: chassis welding on a 1983 BMW E30, in-ear biosensing research at WPI with a patent filing in progress, and FORGE — an AI engineering-analysis platform built solo. Fusion 360, AutoCAD, C++, Python, JavaScript.
One mindset across all of it: understand the system deeply, then build something better.
Away from the bench I climb — lead 5.10d, top rope 5.11c, boulder V7, lead-certified one month after first touching a wall. Fluent in Turkish and English.
Resume
Download PDF
Hikmet Bisen · Austin, TX · 737-351-8670
hikmetbisen@gmail.com ·
linkedin.com/in/hikmetbisen ·
github.com/HikmetBisen ·
hikmetbisen.github.io/Portfolio
Machine learning · Mechanical engineering · Medical devices
Austin, TX · GPA: 4.56 · Concurrent dual enrollment at Austin Community College
College-level coursework in mathematics, physics, and engineering completed during high school
- Programming languages: Python, JavaScript, Java, C++, MATLAB, HTML/CSS
- Frameworks & tools: FastAPI, React, REST APIs, SQLite, Git & GitHub, Supabase (pgvector)
- AI / ML: retrieval-augmented generation (RAG), LLM integration (Claude, GPT-4o, LiteLLM), Voyage AI embeddings, prompt engineering
- CAD & hardware: Fusion 360, AutoCAD, TinkerCAD, Arduino
- Spoken languages: Turkish (native), English (fluent)
Solo founder & full-stack developer
- Designed and built a full-stack AI platform where engineers upload technical drawings and PDFs and receive structured, citation-backed engineering answers; Python/FastAPI backend with a React 18 frontend.
- Engineered a retrieval-augmented generation (RAG) pipeline: PDF text extraction, Voyage AI embeddings (1,536-dimension vectors), and cosine similarity search over a Supabase pgvector document store.
- Implemented multi-model LLM routing with LiteLLM, automatically directing queries between Claude and GPT-4o based on content (e.g., routing stress, FEA, and fatigue calculations to the stronger model); exposed 5 rate-limited REST API endpoints.
- Shipped 6 of 8 planned phases (RAG, API, UI, AI-guided onboarding, structured PDF report export, multi-model routing); designed a tiered SaaS monetization roadmap. Code public on GitHub.
Independent automotive engineering & restoration project
- Built hands-on expertise in mechanical systems, troubleshooting, and precision-based reassembly.
- Disassembled the vehicle to evaluate mechanical and structural condition, removed rust, reinforced compromised chassis areas, and systematically repaired and reassembled key components.
- Launched and led a full ground-up restoration with my brother, conducting technical research, consulting automotive specialists, and sourcing tools and materials.
- Completed a 9-week web programming course covering foundational and intermediate Python concepts.
- Built several technical projects in Python and attended professional development workshops on time management and technical portfolio development.
Medical device R&D intern
- Led the design and development of a wearable earbud to track vital signs (heart rate, respiratory patterns, vocal activity, and motion) by targeting a subaural site near the jaw shown in recent research to yield rich physiological data.
- Collaborated directly with a WPI professor from concept to prototype, using AutoCAD for mechanical design and C++ for embedded programming, prioritizing a lightweight and comfortable form factor.
- Patent filing in progress for the final design; co-authoring a paper for submission to the MIT Undergraduate Research Technology Conference (URTC).
- Explored mechanical, civil, and electrical engineering through workshops, lectures, and real-world simulations.
- Led and collaborated on multidisciplinary design challenges; applied civil and mechanical concepts to design and build functional model roller coasters with structural analysis and material selection.
- Conducted an independent research project analyzing UT Austin's financial structure — revenue sources, profitability, and budget allocation — strengthening financial analysis and data interpretation skills.
- Gained hands-on exposure to architecture and engineering practice by working with startup firms in real-world design and construction settings.