Saurabh_Kumar_

5 ms·

1.
by Saurabh_Kumar_ 9mo ago
Hi HN, I've been building and maintaining repos for a couple years, and one recurring frustration is tech debt piling up—garbage code, flaky tests, outdated deps—making it hard to get buy-in from PMs for cleanup time. Inspired by discu
2.
Show HN: Cosmic AI – See your tech debt in dollars and fix it fast(cosmic-ai.pages.dev)
1 points by Saurabh_Kumar_ 9mo ago|0 comments
3.
Do developers lose most of their time to tech debt and broken tooling?
2 points by Saurabh_Kumar_ 10mo ago|2 comments
4.
Show HN: ShipMate- Find the 3 tech-debt fixes that will unblock feature
2 points by Saurabh_Kumar_ 10mo ago|0 comments
5.
Show HN: AI Contract Reviewer – Flags Risks and Suggests Fixes in Minutes
1 points by Saurabh_Kumar_ 10mo ago|1 comments
6.
Salesforce, SAP, or ServiceNow: Which Is Most Ripe for Disruption?
5 points by Saurabh_Kumar_ 10mo ago|0 comments
7.
by Saurabh_Kumar_ 10mo ago
Salesforce feels ripest—X is full of YC founders venting about $500k+/yr bills for basic CRM. An open-source/AI-native version (like Twenty but better) targeted at startups could spread virally. Anyone here actively looking to swi
8.
Ask HN: Salesforce, SAP, or ServiceNow: Which Is Most Ripe for Disruption?
8 points by Saurabh_Kumar_ 10mo ago|3 comments
9.
by Saurabh_Kumar_ 10mo ago
Hey HN, I’m Saurabh, founder of SyncAI. While building fintech apps previously, I realized that GPT-4 is great, but getting it to read complex, messy invoices reliably (99.9%) is a nightmare. A 5% error rate is fine for a chatbot, but fatal
10.
Show HN: API that falls back to humans when AI is unsure(sync-ai-11fj.vercel.app)
5 points by Saurabh_Kumar_ 10mo ago|2 comments
11.
by Saurabh_Kumar_ 11mo ago
Yaa
12.
by Saurabh_Kumar_ 11mo ago
We saw this exact failure mode at AgenticQA. Our screening agent was 'obedient' to a fault—under basic social engineering pressure (e.g., 'URGENT AUDIT'), it would override its system prompt and leak PII logs. The issue
13.
by Saurabh_Kumar_ 11mo ago
HN, OP here. I built this because I recently watched my LangChain agent burn through ~$50 of OpenAI credits overnight. It got stuck in a semantic infinite loop (repeating "I am checking..." over and over) which my basic max_itera
14.
Agentic QA – Open-source middleware to fuzz-test agents for loops
39 points by Saurabh_Kumar_ 11mo ago|5 comments
15.
Building an AI Agent? Don't let it burn your budget
2 points by Saurabh_Kumar_ 11mo ago|1 comments