Sambit Mahapatra | Construction Compliance AI (@sambit9238) header
Sambit Mahapatra | Construction Compliance AI (@sambit9238) profile picture

Building private AI for construction compliance Documents → cited obligations → evidence → defensible closeout

Bengaluru
Joined May 2014
Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)Photo by Sambit Mahapatra | Construction Compliance AI (@sambit9238)
Sambit Mahapatra | Construction Compliance AI retweeted
We're giving away a CONTROL Resonant custom wrapped NVIDIA GeForce RTX 5080 to celebrate its official release with #RTXON. To enter: 🟢 Share this post 🟢 Comment #RTXON T&Cs: nvda.ws/4z24Gcq
37,018
30,532
2,813
33,141
3,114,984
Sambit Mahapatra | Construction Compliance AI retweeted
NVIDIA released a fast speaker-labeling model today: Nemotron 3 Diarization. Like all of the Nemotron family from @NVIDIA_AI, this is an open weights model, so you can grab the weights from @huggingface and use the model in the cloud or on your local device. Nemotron 3 Diarization is a useful building block for voice agents and realtime AI applications that need to handle speech from multiple people at the same time. It handles overlapping speech well, can identify up to eight speakers, and pairs with any transcription model you're using in your multi-modal pipelines. @jonptaylor built a demo (code below) and recorded a technical deep-dive showing how to use this diarization model together with a speech-to-text model. Jon's code also uses Jev for super-accurate confirmation of critical voice input in a noisy/multi-speaker environment.
13
13
9
183
12,947
Sambit Mahapatra | Construction Compliance AI retweeted
AGENT HARNESSES ARE THE NEW GPT WRAPPERS What exactly is an agent harness? A harness does 4 things: 1. Runs the model in a loop so it keeps working step after step instead of answering once and stopping. 2. Gives it hands to read files, call tools, open portals, and run code. 3. Manages its memory so hour three of a job still knows what happened on hour one. 4. Enforces the rules about what it can touch and when it has to stop and ask a human. WHY IT'S INTERESTING - The market is MEGA. I think about it like a wrapper let you sell software ($800B market) but a harness lets you sell the work ($5T+ market)! - It doesn't depend on any one company's model. The knowledge about the job lives in the harness, so you can run GPT today, Claude next month, an open model like Google Gemma, Qwen, Deepseek etc on your own machine after that, and it keeps working. A wrapper was one model doing everything and a harness is a router. - It gets better the more you use it. See, every correction a human makes becomes a rule the harness keeps, so what you have after 500 jobs is a completely different product than what you had after 5. Wrappers only got better when OpenAI got better. - You can charge for finished work! The harness knows when a job is done, so you can price per claim, per filing, per review, per resolved exception, or per closed month. This helps compete with SaaS! - Most jobs are just the same handful of decisions, repeated, using the same few tools. That's true for most of the 800+ occupations out there, which is why almost all of them could have a harness built for them. Lots of space for startups to be building. -The frontier labs won't come for these. OpenAI is not going to learn how a freight claim gets denied in Rotterdam or which prior auth your specific payer rejects. Those markets are too small for them and the knowledge only exists inside the job. They'll keep making the models better, which just makes your harness better which is cool. - OH, AND really important, you can ACTUALLY build one now. OpenAI's Agents API, Claude's managed agents, Vercel's eve, LangChain harness framework etc. They all make it way easier to build agent harnesses. Agent harnesses truly are the new GPT wrappers. This is just the beginning.
238
229
58
2,486
182,023
Sambit Mahapatra | Construction Compliance AI retweeted
Either you die a system of record or you live long enough to become a domain-specific harness
I was just at YC demo day yesterday. Besides hardware/physical things, everyone is just basically just building a domain-specific harness
193
292
140
4,439
1,459,947
Sambit Mahapatra | Construction Compliance AI retweeted
Best local AI models to run on your DGX Sparks ✨ 1x DGX Spark - Qwen3.8-Flash-Next github.com/MiaAI-Lab/Qwen3.8… ------------------ 2x DGX Sparks DeepSeek v4 Flash github.com/MiaAI-Lab/DeepSee… GLM 5.3 Flash github.com/MiaAI-Lab/GLM-5.3… Qwen3.8 Flash github.com/MiaAI-Lab/Qwen3.8… ------------------ 3x DGX Sparks - DeepSeek v4.1 Flash TP3 github.com/MiaAI-Lab/DeepSee… ------------------ 4x DGX Sparks - DeepSeek v4.1 Flash TP4 github.com/MiaAI-Lab/DeepSee…
52
63
18
598
87,699
Sambit Mahapatra | Construction Compliance AI retweeted
The Nebius AI Builder Program is now available. AI isn’t just a model you call anymore. It’s a system you build. And builders, not a few closed labs, will decide what it becomes. The open ecosystem has all the pieces. We want to make it easier to put them together and start building. The program is free, with $400+ in credits and discounts, working code and cookbooks, office hours with engineers, and a community to build with. We’re joined by @NVIDIAAI, @LangChain, @huggingface, @cognition, @OpenHandsDev, @tavilyai, @TolokaAI, @composio, @PrimeIntellect, @MiniMax_AI, @Alibaba_Qwen, and more joining soon. Join with the link in the comments 👇
479
385
269
1,229
2,607,365
Sambit Mahapatra | Construction Compliance AI retweeted
RT And Like This Post To Let @Alibaba_Qwen Know We Appreciate everything they've done. But that Qwen 3.8 27B would like to have it's Soul Mate 3.8 35B A3 as well. It is only right.
12
47
334
8,644
Sambit Mahapatra | Construction Compliance AI retweeted
We built the fastest PDF -> markdown parser in the world 🚀⚡️ AND it’s more accurate than any other open-source, model-free parser (pymupdf4llm, opendataloader, pdf-inspector, markitdown) on 3 standardized benchmarks: olmOCR0-bench, opendataloader-bench, ParseBench Introducing LiteParse v2.1. The v2 base version was already the fastest document->text parser on the planet, and with this new release we’ve introduced markdown. It is fully open-source (Apache 2.0) and free, is usable from CLI/Rust/Node/Python/WASM, and is also installable as a one-click agent skill. Check it out: llamaindex.ai/blog/markdown-… Come check out LiteParse: github.com/run-llama/litepar…
LiteParse v2.1 is here, and its bringing the fastest markdown output possible. In this release, we are fulfilling our top request: markdown output. But in the spirit of "lite"-ness, we are doing this completely LLM-free and fast. Not only is it fast, it also beats all other model-free competitors in 3 separate benchmark datasets. Read more about it in our release blog: llamaindex.ai/blog/markdown-…
73
199
18
1,633
364,903
Sambit Mahapatra | Construction Compliance AI retweeted
What if you could build a working AI agent in just 4 lines of code? LangChain and LangGraph are powerful, but they also come with concepts like chains, runnables, and state graphs. That is useful for complex workflows, but overkill when you just want to give a model a tool and let it solve a task. smolagents from Hugging Face is a barebones agent library designed for simplicity. In roughly 1,000 lines of code, it gives you the core pieces needed to build a working agent with a single import and three lines of setup. Key features: • Agents act by writing Python in a sandbox, the same way you'd script a task in a notebook • Model-agnostic design, swap between OpenAI, Anthropic, local Ollama, or Hugging Face Inference • Push agents to the Hugging Face Hub to share with the community --- 📬 I share 2 practical tips on practical tools for data and AI twice a week on Substack. Subscribe here: bit.ly/46fdOPl #AI #Python #LLM #HuggingFace
1
1
5
331
Sambit Mahapatra | Construction Compliance AI retweeted
Composer 2.5 is now live For the next week, we’re doubling the included usage
Introducing Composer 2.5, our most powerful model yet. It's more intelligent, better at sustained work on long-running tasks, and more reliable at following complex instructions. For the next week, we’re doubling the included usage of the model.
44
30
4
816
45,239
Sambit Mahapatra | Construction Compliance AI retweeted
If Hermes doesn't work when the MacBook lid is closed, Then I figure out a solution. run : 𝚜𝚞𝚍𝚘 𝚙𝚖𝚜𝚎𝚝 -𝚋 𝚍𝚒𝚜𝚊𝚋𝚕𝚎𝚜𝚕𝚎𝚎𝚙 𝟷 Voilà.
14
6
1
184
9,237
Sambit Mahapatra | Construction Compliance AI retweeted
PyMuPDF4LLM now includes Layout: GNN-based document layout analysis that runs on CPU. Also new: pluggable OCR with auto-detection. Tesseract, RapidOCR, PaddleOCR supported, or bring your own. Try the demo → demo.pymupdf.io
4
4
237
Sambit Mahapatra | Construction Compliance AI retweeted
This is Claude Sonnet 4.6: our most capable Sonnet model yet. It’s a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design. It also features a 1M token context window in beta.
1,067
2,419
1,568
21,760
7,613,416
Sambit Mahapatra | Construction Compliance AI retweeted
31
305
10
2,857
42,290
Sambit Mahapatra | Construction Compliance AI retweeted
I found a way to sell AI infrastructure to companies doing $2M-$50M/year. ALL with no sales team, enterprise connections or cold calls. Businesses are desperate for Ai implementation right now and anyone can do this. But most people get it wrong and that's why they fail. They lead with tools, ex) "we build automations" or "we integrate AI." That means nothing to a CEO managing 40 people and $5M in revenue. They pitch features instead of showing the cost of doing nothing. And they price hourly, so the buyer treats them like a freelancer instead of a partner. Mid-market and enterprise companies are bleeding $100K-$500K/year on broken processes, bloated SaaS stacks, and manual work they don't even realize they're paying for. They WANT to buy AI infrastructure, they just don't trust most people selling it. Because most people selling it sound like every other agency. I created a guide breaking down exactly: → How to position AI infrastructure so executives actually listen → The discovery framework that turns a 30-min call into a $25K-$100K project → How to calculate ROI so the price sells itself → The 3-pillar strategy process that closes 60%+ of qualified prospects → Why "sell the map before you sell the build" changes everything RT + reply "INFRA" and I'll send you the FULL guide (must follow so I can DM)
960
397
5
735
76,462
Sambit Mahapatra | Construction Compliance AI retweeted
6,522
27,542
13,444
117,781
88,196,666
Sambit Mahapatra | Construction Compliance AI retweeted
Avatar: Fire and Ash opened to $345M worldwide in its first weekend, third biggest opening of the year. The film reportedly carries a budget north of $400M. For context, Avatar: The Way of Water debuted at $435M globally.
70
164
19
5,466
749,957
Sambit Mahapatra | Construction Compliance AI retweeted
If I can train someone to do your job in a week, you will always be paid the minimum. If your specific knowledge takes 10 years to acquire, you can name your price. Build a moat around your brain.
142
122
7
1,438
37,133
Sambit Mahapatra | Construction Compliance AI retweeted
I’m low-key shocked this list is still sitting on my desk, so… here we go. I’ve got 6,000+ DECISION MAKERS ACROSS THE INDUSTRIAL BACKBONE SECTORS: - MACHINERY - MANUFACTURING - BUILDING MATERIALS - PACKAGING - ENGINEERING & TECHNICAL SERVICES These are actual buyers; people who sign off on equipment, approve vendor shifts, green-light capex, and hold the budget. This dataset is stacked with direct work emails, job titles filtered to authority roles, company segmentation by size & revenue, NAICS-mapped industries, cleaned, de-duped, and ready to plug into your outbound tools. If you sell to: - Automation tools - Industrial SaaS - Heavy machinery - Maintenance services - Packaging solutions - Safety equipment - Energy optimization tech - Or anything that touches factory floors, supply chains, or production systems… …this list will hit the heart of your ICP without you burning hours on filtering. Want it? Comment "171" + Like + Repost and I'll DM it to you. (must follow for DM)
58
36
75
4,945