Minh Vu
Software Engineer · M.Sc. student moving into low-level & systems
I'm a software engineer and a Computer Science master's student at the University of Passau. Now I'm moving into low-level and systems work, like compilers (especially LLVM and MLIR), database engines and distributed systems.
I have 3+ years of industry experience in backend, applied AI and the Elastic Stack. At Coder Trove, I was the first engineer on the Grapple AI team. Grapple AI is an Australian lender's platform for checking borrowers' documents. I built its AI backend from an empty repository. At Flowzar in Munich, I built a self-serve upload that replaced a manual import.
I started with competitive programming. I won Third Prize at the ICPC Asia Hanoi Regional (ranked 32/126) and Second Prize at the Vietnam Olympiad in Informatics for Undergraduates. My LeetCode rating is 1756 ↗ (top 10.4%, Sep 2026).
Experience
AI Software Engineer (Working Student) · Flowzar GmbH
- Built a self-serve CSV/Excel upload for a cash-flow forecasting product, with files of up to 10,000 rows. Before this, the principal engineer had to redo the import for every new file a customer sent. Column mapping uses fixed rules first and an LLM only when the rules fail. The data goes into an append-only PostgreSQL ledger. We used it to onboard a major customer.
- Wrote a rule-based transaction categorizer that the principal engineer used for imports. Later I built an LLM pipeline that categorizes transactions automatically.
- Built React features such as a streaming AI chat that shows the assistant's progress while it answers, and a paginated transaction table.
Software Engineer · Coder Trove · Grapple AI ↗
- Full-time from Jul 2024 to Sep 2025; back remote with flexible hours since Jun 2026. Joined as the first engineer on the AI team, which grew to 4 people.
- Built the AI backend from an empty repository (FastAPI on AWS), including the first 4 of its 10 LangGraph assessment workflows: company checks, bulk company checks, invoices and contracts.
- Added a timeout and a backup chain across LLM providers within a day, after our main provider slowed to 10-minute responses. Then I split long contracts into 5-page batches that run in parallel.
- Built a benchmark on a fixed set of 18 documents to measure accuracy, cost and speed before we changed the extraction method. On that set, field accuracy for proof-of-delivery documents went from 78.9% to 100% (4 documents).
- Moved all 3 document-upload flows to direct-to-S3 uploads with presigned URLs, removing two proxy hops and up to about 250 MB of in-process file buffering per upload batch.
- Found why some assessments showed the wrong company. The team suspected the LLM, but the cause was a cache key that ignored the company-number argument. A teammate's fix then made ID lookups match exactly.
Software Engineer · ENTIIS Pte Ltd
- Wrote Logstash parsing scripts for 5 client projects at this Elastic Stack consultancy, one at a Vietnamese bank and four in Singapore. They turned raw logs into structured fields. I also wrote Filebeat configs for shipping logs.
- Built 7 Kibana visualizations on-site at the bank to track latency, uptime and errors across its web, mobile and internal systems. I also made a Kibana Canvas status board that changes color at set thresholds.
workingstudentjobs.de ↗, a job board in English and German for working student and internship jobs in Germany.
As of Oct 2026, it has about 46,000 sessions a month and 950+ email-alert subscribers. I built the pipeline that collects jobs from job sites, ATS APIs and company career pages. It removes duplicates, classifies the jobs with an LLM and publishes a median of about 430 of them every weekday. When DeepSeek raised its peak-hour prices in Aug 2026, I tested an OpenAI model on 60 published jobs. It made the same accept/reject call on 58 of them, so I switched. At peak hours, the LLM cost went from $1.57 to $0.67 per 1,000 jobs. Since Oct 2026, Claude Sonnet 5.5 classifies first, paid by a startup credit, with the DeepSeek and OpenAI models behind it. Errors go to PostHog and every cron job sends a heartbeat to Uptime Kuma.
I started it in March 2026, and it went live in April 2026.
Education
Master of Science in Computer Science
I completed Software Analysis, which covered dataflow analysis and abstract interpretation, and Practical Parallel Programming, which covered MPI, OpenMP, GPU programming and the polyhedral model.
CourseworkBachelor of Information Technology
Projects
- Wrote constant folding, common subexpression elimination and dead-code elimination passes for a custom torch.compile backend. They run until the graph stops changing.
- Fixed bugs where in-place writes, also through tensor views, made the passes give wrong results, and wrote regression tests for four of them. An end-to-end test compares the output with plain PyTorch.
Formularizer
- Built an AI tool that turned plain-English requests into spreadsheet formulas, and grew it to 20,000+ users and 2,000+ monthly active users.
- Routed requests across several LLM providers to keep costs down, with automatic fallback when one failed.
- Served 10,000+ requests a month at peak, with payments through Stripe.
Honors & Awards
Third Prize, The 2021 ICPC Asia Hanoi Regional Contest (ranked 32/126)
Second Prize, Vietnam Olympiad in Informatics for Undergraduates
Top 10 of Graduating Batch of 900+ students, FPT University
Skills
Languages
From completed M.Sc. coursework
Backend
AWS
Databases
Infrastructure
Monitoring
LLM applications
I'm looking for a working student job or a 2027 software engineering internship, in Munich or anywhere in Germany. I'm most interested in low-level and systems work, like compilers, databases and distributed systems.
Let's Connect