HalfStack Vienna 2026
•
Nov 19, 2026
The Feature That Didn’t Need to Be Smart
Carmen Huidobro
- Discuss this with your agent
- Open in Claude
- Open in ChatGPT
How do you build a genuinely useful LLM feature without relying on cloud APIs or sacrificing user privacy? See how a skeptical developer tamed local open-weight models using strict JSON validation and zero-temperature retries.
Matching moments
More from HalfStack Vienna 2026
Upcoming sessions on this topic
Starts in 6 days
Conference India 2026
November 25–26, 2026
Bengaluru, India
-
Your AI Agents Are Bleeding Tokens: Building a Runtime Cost Control Plane for LLMs

Aakriti Aggarwal & Bhavishya Pandit
Wed 25 Nov · Stage 1 - powered by IBM
Related videos
Related articles
CH
Chris Heilmann
CH
Chris Heilmann
BB
Benedikt Bischof
LM
Luis Minvielle
From learning to earning
Jobs that call for the skills explored in this talk.
about 2 months ago
•
AI Software Engineer (Germany)
Sunhat
Berlin, Germany
Expert
€70k–100k
Remote
REST
TypeScript
AI Frameworks
9 days ago
Principal Software Engineer, Copilot Code Review
GitHub
Beverly Hills, CA, United States
Expert
Remote
Ruby
Golang
Github
4 days ago
Staff Software Engineer, Copilot Code Review
GitHub
Beverly Hills, CA, United States
Experienced
Remote
Ruby
Golang
Github
2 months ago
•
AI & Machine Learning Engineer (all genders)
msg
Frankfurt am Main, Germany
Experienced
Hybrid · 3 days on-site
Python
Langchain
Machine Learning
20 days ago
Machine Learning Engineer 2
Twilio
San Francisco, CA, United States
$128k–160k
Remote
Twilio
Backend
Raw Data
16 days ago
•
Senior Data & AI Architect
reeeliance IM GmbH
Berlin, Germany
Expert
Hybrid · 3 days on-site · Remote
Python
Microsoft SQL Server
Cloud (AWS/Google/Azure)