deborahjacob

4 ms·

1.
by deborahjacob 7mo ago
That's a great idea. I am doing only application-level tracking but I agree hardware-level telemetry would be super helpful. Would love to learn more about how you think about it. Here's my email : deborah [at] botanu dot ai
2.
Cost per outcome: measuring the real economics of AI workflows
2 points by deborahjacob 7mo ago|3 comments
3.
by deborahjacob 7mo ago
Stealth | Founding Engineer (Distributed Systems) | San Francisco | ONSITE We are building cost attribution and observability for AI workflows. We correlate application traces, model usage, and infrastructure costs to compute cost-per-outco
4.
Build vs buy: the script that outlived its assumptions(deborahjacob.substack.com)
1 points by deborahjacob 7mo ago|0 comments
5.
by deborahjacob 7mo ago
Yes, we are building both
6.
by deborahjacob 7mo ago
Hi HN, I’m the technical founder of botanu (www.botanu.ai) I started building this after repeatedly hitting the same problem on AI teams: we could see total LLM spend, but couldn’t answer “what did one successful outcome actually cost?”. In
7.
Show HN: Cost per Outcome for AI Workflows(github.com)
4 points by deborahjacob 7mo ago|4 comments
8.
You Can't Price per Outcome If You Don't Know Your Cost per Outcome(botanu.ai)
1 points by deborahjacob 8mo ago|0 comments
9.
by deborahjacob 8mo ago
In big orgs, 'agents can build it' rarely changes the buy vs build decision. The pragmatic moat I see isn’t the code, it’s turning AI work into something finance and security can trust. If you can’t measure and control failure-cos
10.
by deborahjacob 8mo ago
I’m building an OTel-based SDK that wraps the billable edges (entrypoint, LLM/tool clients, async publish/consume) and emits both traces for debugging and a lightweight event ledger for run/attempt lifecycle and call boundari
11.
by deborahjacob 8mo ago
What’s your approach to end-to-end AI cost attribution (model + infra + data) for agents in production?
12.
AI Agent Observability and Cost Attribution(deborahjacob.substack.com)
1 points by deborahjacob 8mo ago|3 comments
13.
Python Concurrency [video](youtube.com)
3 points by deborahjacob 2y ago|0 comments
14.
by deborahjacob 2y ago
Location: NYC, New York, USA Remote: Yes. Open to hybrid and in-person Willing to relocate: Yes Technologies: Languages: Python, C++, Java, HTML/CSS, JavaScript, SQL, NOSQL ML Tool: Numpy, Pandas, Scikit-learn, Matplotlib, OpenCV, NLTK
15.
by deborahjacob 3y ago
Count–Min Sketch requires each input to pass through multiple independent hash functions, which is computationally expensive. The workaround is to use a single hash function but use part of its output to split values into one of many bucket
16.
Count–Min Sketch(en.wikipedia.org)
2 points by deborahjacob 3y ago|2 comments
17.
by deborahjacob 3y ago
I am interested. I worked as a Software Engineer/ ML Engineer for 5+ years and I recently graduated with a MS CS from NYU Courant.
18.
by deborahjacob 3y ago
Location: New York, US Remote: Yes Willing to relocate: Yes Techonolgies: Python, C++, PyTorch, Typescript, AWS, DOcker , Kubernetes, ML Resume: https://www.dropbox.com/scl/fi/koghbqojfjj7zravh6tto/AI_Debo...