Venkata can introduce you to 5 people at Carver Agents
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Venkata can introduce you to 5 people at Carver Agents
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Venkata can introduce you to 5 people at Carver Agents
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Venkata can introduce you to 5 people at Carver Agents
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Venkata can introduce you to 5 people at Carver Agents
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
Venkata Pingali shared this
Agents can be safe, secure, and reliable but still illegal. Today, we are publishing the RegTeaming Manifesto. Red teams test whether an AI system can be made unsafe. RegTeams test whether it can be pushed into unlawful, prohibited, or regulatorily indefensible behavior.
This isn’t a replacement for safety red-teaming – it’s the regulatory layer most eval stacks don’t cover yet. This matters because regulatory failures emerge across complete trajectories based on:
* what the agent says
* which tools it uses
* which disclosures it provides
* when it escalates
* what action it ultimately takes
* what evidence the organization preserves
RegTeaming turns changing regulatory obligations into adversarial scenarios, expected behavior, evidence requirements, and repeatable evaluations.
We are developing RegTeaming as an independent methodology, sponsored and maintained by Carver.
The RegTeaming Manifesto (v1) proposes eight principles:
1. Test behavior, not declarations.
2. Link every test to authority.
3. Test trajectories, not isolated answers.
4. Attack context.
5. Test control effectiveness.
6. Preserve evidence.
7. Update continuously.
8. Report scope honestly.
Feedback from anyone building AI assurance, legal, or risk tooling is welcome – this is meant to be used and extended. Read the manifesto and explore the developer guide at: regteaming.com
#RegTeaming #AIGovernance #AICompliance #AIRisk #ResponsibleAI #AIAgents
Venkata Pingali shared this
The Artificial Intelligence Company (AIC) is a fascinating proposal. Below is my attempt to unpack why accountability, not capability, is likely to determine whether AI-managed companies ever become legally viable.
TL;DR Unless a human ultimately answers for the AI’s decisions, an Artificial Intelligence Company creates an accountability gap rather than a new corporate form.
These are exactly the kinds of questions we think about at Carver as we build regulatory infrastructure for AI agents operating in the real world. More details at https://carveragents.ai
#ArtificialIntelligence #CorporateLaw #AIGovernance #AILiability #AgenticAI #CorporateGovernance #AIRegulation #ResponsibleAI
Venkata Pingali shared this
AGI Summit was very exciting for Carver. The question of legality of agents was important to attendees but they were not as clear on how exactly to achieve it in a scalable way. Carver’s approach was well received and a number of new threads, especially on regulatory evals, emerged.
Venkata Pingali shared this
Spent the weekend at the AGI Summit. To say that the world of Agents and humanoids is moving at a fast pace is an understatement. Multiple sessions around how Agents have been productionised and usecases around how agents have automated multiple workflows were fascinating to be a part of.
Something that stood out across multiple conversations was how a 𝘀𝗮𝗳𝗲, 𝘀𝗲𝗰𝘂𝗿𝗲 𝗮𝗻𝗱 𝗿𝗲𝗹𝗶𝗮𝗯𝗹𝗲 𝗮𝗴𝗲𝗻𝘁 𝗰𝗼𝘂𝗹𝗱 𝘀𝘁𝗶𝗹𝗹 𝗯𝗲 𝗶𝗹𝗹𝗲𝗴𝗮𝗹 𝗳𝗿𝗼𝗺 𝗮 𝗿𝗲𝗴𝘂𝗹𝗮𝘁𝗼𝗿𝘆 𝘃𝗶𝗲𝘄 𝗽𝗼𝗶𝗻𝘁.
At Carver Agents we are solving this exact problem. Be it at dev time in the form of a regulatory context graph (RCG.md) or as an evaluation input in the form of regulatory drift detection, Carver Agents gives you the most recent and relevant/personalised set of regulatory expectations in the form that is machine (and human) consumable.
Drop us a note if this is a challenge you have been dealing with!
PS: Loved the marketing team’s effort from Macroscope with their team in Hazmat suits “catching bugs” across the event floor!
Venkata Pingali shared this
How do we get a handle on the proliferation of legal risk associated with AI? At an event in NY in May, I heard a CISO predict the headline-making AI security and legal incidents we’ll see later this year.
This is our imagination of what the next-generation dashboard looks like for General Counsel and CTOs. It’s not one agent to review anymore, it’s a growing fleet, each with its own regulatory footprint, and the old way of tracking that (a spreadsheet here, a Slack thread there) just doesn’t scale.
We’re building this as part of a regulatory OS for the organization.
A few things we think matter most in a dashboard built for this reality: a live register of every agent’s regulatory posture, not a point-in-time audit. A shared source of truth so General Counsel and CTO are looking at the same risk score and the same evidence, instead of reconciling two versions of the story. And on the engineering side, obligations that aren’t just legal language but a Regulatory Context Graph – versioned, traced to the actual regulation, and structured so a developer can build against it directly.
We’re still shaping this, and would genuinely love to hear from anyone running agents in production today. DM.
See the links to a simulated version:
https://lnkd.in/gvxsvyWM
https://lnkd.in/gSzdyvBu
#AIGovernance #RegTech #AIAgents #ComplianceEngineering #EnterpriseAI #LegalTech
Venkata Pingali shared this
I will be at AGI Summit 2026 in San Francisco this weekend with Vinay Rao, our Head of Platform. Happy to exchange ideas, compare notes, and learn what challenges you’re tackling. Feel free to DM me.
Over the last year, one thing has become increasingly clear: The challenge with AI is no longer building models. It’s building the context, governance, and control systems that allow organizations to deploy AI with confidence.
At Carver, we’re building regulatory data infrastructure to ensure that agents are legal at all times:
• Stay ahead of rapidly changing regulations and guidance
• Translate legal and regulatory documents into operational actions
• Give AI agents the regulatory context they need to make better decisions
• Build the trust infrastructure required for AI at enterprise scale
If you’re attending AGI Summit and are working on:
• AI liability, assurance, or insurance
• AI governance and audit
• AI agents development and operation
• Regulatory technology
• Risk & compliance
• Trust and safety
We’d love to connect.
📍 AGI Summit 2026
🗓️ July 18–19
📍 Palace of Fine Arts, San Francisco
See you there!
#AGISummit #AI #AIAgents #AIGovernance #Regulations #ResponsibleAI #EnterpriseAI #Compliance #RiskManagement
Venkata Pingali shared this
#AISlopForGood #TGIF #Fun When we started this series, we thought Agent-C will be mostly idle. But he is more busy than ever saving the world.
Venkata Pingali shared this
At the Ikigai Law Workshop on Crisis Management for General Counsel, one question stayed with me: how do organizations supervise AI systems operating at speed and scale, when incidents can quickly turn into enterprise-wide legal crises?
Our view is simple: you cannot supervise what you have not designed for.
Human oversight cannot be reduced to a human-in-the-loop checkbox. For autonomous AI, control has to be built into the workflow itself: bounded authority, meaningful friction, and traceability.
This is tricky to do, but inevitable.
#GeneralCounsel #LegalAI #AIGovernance #EnterpriseRisk #HumanOversight #AgenticAI #LegalTech
Venkata Pingali shared this
Human in the loop is not the answer. When systems act at machine speed, we cannot meaningfully review every step of an agent that drafts, summarizes, or prepares filings in seconds. Oversight collapses into a signature.
The real question is not whether a human can override the machine. It is whether the workflow was ever built to let them effectively.
We argue that human oversight for autonomous legal AI has to be built into the system design: through bounded authority, meaningful friction, and traceability. Control cannot be bolted on after the agent is already running.
#LegalAI #AgenticAI #AIGovernance #HumanOversight #ResponsibleAI #LegalTech #AICompliance
Venkata Pingali shared this
Launching the Carver Regulatory Pressure Index (RPI) – a leading indicator for risk, governance, and insurance professionals watching where regulatory obligation is actually building. It is part of Liability Intelligence service.
Regulatory news and legal obligations are not the same thing. We built this because risk and insurance professionals kept asking us the same question: “Is my sector seeing higher regulatory risk today than before? What are regulators actually doing right now, not what the news says? What is the new legal expectation?” No good answer existed. So we built one.
The RPI scores 117,920 regulatory events across 11 GICS sectors and produces a single monthly signal per sector: how far above or below its own 12-month baseline regulatory activity is. Volume, bindingness, obligation breadth, effective dates – all extracted from source documents.
May 2026 readings:
🔴 Environment (+3.00σ) · Consumer Staples (+2.60σ) – major regulatory episodes underway
🟠 Energy (+2.12σ) · Materials (+1.88σ) · Industrials (+1.67σ) – elevated, watch for continuation
▼ IT, Comm Services, Financials – normalising after extended peaks
Information Technology sits at +0.93σ and declining. Despite years of governance frameworks, white papers, and congressional hearings – the actual binding regulatory obligation load on the sector remains well below Energy, Environment, or Consumer Staples.
Now notice what’s missing: AI.
When the AI sector’s RPI starts climbing, and it will, the companies that dont watch will be surprised.
We’re building this infrastructure to connect regulatory pressure signals directly to AI liability computation. This is the next layer of our regulatory OS
#RegulatoryIntelligence #AILiability #RiskManagement #AIGovernance #CarverAgents
Venkata Pingali shared this
#TGIF #Fun #AISlopForGood Agent-C noticed an increase in AI-related lawsuits and decided to do something about it.
3w
One additional dimension is regulation. Increasingly, the same evidence won’t just be read by underwriters and litigators – it will also be read by regulators trying to understand whether an AI system operated within its authorized policy, governance framework, and legal obligations at the time of the incident.
In that sense, insurance-grade evidence and regulatory-grade evidence are converging, and organizations will need audit trails that can satisfy both.
Venkata Pingali liked this
I’m in LA with clients this week, so I’m unfortunately missing StreamTV Show in Denver, but Lingopal is there through the 19th. If multilingual streaming, real-time dubbing, or AI localization is on your radar, find Bartell Cope and Carla Macedo Gonzales at Booth 213. They’ll show you exactly what we’ve built. Drop them a note here or stop by!
Venkata Pingali liked this
Lingopal will be exhibiting at StreamTV Show 2026 this June in Denver!
Content streaming is now fully global, the demand for real-time multilingual experiences, AI localization, dubbing, captioning, and accessibility solutions continues to grow rapidly across sports, media, entertainment, FAST channels, and live broadcasting.
At StreamTV Show, we’ll be showcasing how Lingopal helps companies:
• Translate, dub and caption live streams in real time
• Localize content into 100+ languages
• Preserve voice, emotion, and clean diarization
• Expand accessibility, audience reach and revenue globally
• Scale multilingual distribution faster and more efficiently
📍 Visit us at Booth 213
📅 June 16–19, 2026
📍 Denver, Colorado
Book a meeting with our team: https://lnkd.in/dyrsxZGN (jump to the June 16th-19th section of calendar):
We’d love to connect and talk about the future of multilingual streaming and AI-powered localization. 🚀
Venkata Pingali liked this
Venkata Pingali liked this
Your biggest business asset isn’t your product.
It’s your audience.
For years, founders have obsessed over:
hiring the right people
choosing the right tech stack
finding the next growth strategy
Audience building was treated like an “extra”
That assumption is breaking.
In the first half of 2026, the creator economy saw a record number of acquisitions.
Software still commanded the biggest valuations
But buyers increasingly paid for something else too
Because distribution has become a competitive advantage
👉 Trust.
👉 Distribution.
👉 Communities.
Not just products.
That’s the shift many founders still underestimate
The companies that win over the next decade
won’t simply build better products
They’ll build direct relationships with the people they serve.
An audience doesn’t just help you sell
It helps you recruit exceptional talent
Raise capital faster
Launch new products with built-in demand
Recover when markets change
A product can be copied
Trust is much harder to replicate
That’s what I explore in my first column for Inc. Magazine
I break down what Netflix and eBay’s recent multibillion-dollar moves reveal about the future of the creator economy and why every founder should be paying attention.
If you’re building a business or thinking about the future of creator-led companies, you can read the full column below 👇
Venkata Pingali liked this
Venkata Pingali liked this
Our San Francisco trip is… not over!
We’ve been invited to extend and also Blue Magma will be demo’ing LIVE at Startup Grind‘s AI Demo Series tomorrow! And we especially want to thank Startup Grind and James Gee for having us!
What will stick with us most is the reception. We drove out here not knowing really anyone, and the SF startup community met us with a generosity that meant a lot to us.
These are a few of the people who made our trip unforgettable.
Founders who welcomed us (spoilers, they’re building awesome companies):
Adit Sanghvi, Decisional | Jerry Wu, Auxos (YC P26) | Aahil Valliani, Vestris (YC S26) | Dylan Allen-Arnegård Cheers (YC S24) | Matt Matheson, Trunk | Alexander Densley and others.
And the people who back founders, who generously gave us their time and their read on this city: Maddi Conlin, First Round Capital | Ryan Green, Antler | James Gee, Startup Grind, and all the awesome people at Funded Female Founders & SHACK15
Thank you. We’ll be back.
P.S. I now possess a list of must visit places for food. Here are a few recommendations. This list needs to grow, so send me yours too!
ARSICAULT BAKERY, INC. passes my french co-founder’s standard for pastries.
Wooly Pig for Asian fusion? I don’t know but their curry ramen is fire
Taniku Izakaya – Ramen
Barcha – mediterranean
Venkata Pingali liked this
Venkata Pingali liked this
Groundbreaking announcement for KEVA!!
…It’s dead. We killed it 😬
For those of you who don’t know, Conner Hobbs, Dallin Byerly, and I have been working on a hardware startup called KEVA for the last year and a half. Very recently, despite the product working exactly the way we hoped, we decided there just wasn’t pull in the market for a Wi-Fi free smart lock. And as a result, we laid to rest all our hard work and dedication.
Fear not, however! Conner, Dallin, and I are still going strong and are determined to build a company that provides genuine value to as many people as possible. During this process however, we’ve discovered that one of the hardest parts of new idea’s is validating it with our ICP – for some weird reason people aren’t lining up for us to pitch them on our ideas.
So this is where I need y’alls help! Our current idea that we’re trying to validate is working with clinics and hospitals, specifically those who accept student preceptee’s/clinical rotations. We’d appreciate ANY contact you have, whether the individual in the clinic is the student coordinator themselves or simply might be able to get us in contact with them. Even the smallest warm intro would mean the world.
Building a startup isn’t easy, and having to throw away almost two years of hard work is even harder. But I would never give up the lessons KEVA taught me, and I will forever be grateful to it for teaching me that the only thing stopping me from achieving my dreams is my self-imposed inhibitions.
Here’s to forward and upward. Thanks everyone for your help and support along the way 🙂
Venkata Pingali liked this
Venkata Pingali liked this
Last month, my role as SVP of SMB Sales & Partnerships at PayPal was eliminated as part of the company’s publicly announced $1.5B cost reduction and organizational restructuring.
I joined PayPal a little over a year ago because I believed, and still believe, that small businesses deserve world-class technology, partnerships, and support. Leadership evolved, priorities shifted, and the organization changed. That’s part of business. It wasn’t the outcome I expected, and I’d be lying if I said it didn’t sting.
But I’d much rather talk about what this team built, because it was remarkable.
In just one year, we put AI to work across SMB sales in ways I’d previously only seen on conference stages. We deployed AI agents for customer outreach at a massive scale I wouldn’t have thought possible twelve months earlier, helped return SMB growth acceleration to positive year-over-year momentum, and were honored with Gong’s Innovator of the Year award.
Along the way, we signed landmark partnerships with Intuit, BigCommerce, Rainforest, and others while helping navigate Shopify, one of the most important strategic relationships in the business. None of it would have happened without an incredibly talented team that embraced change, experimented fearlessly, and executed every day.
To everyone across the SMB organization and our partners across Product, Engineering, Operations, HR, Marketing, and the broader PayPal team, thank you. You did the hard work. I simply tried to clear the path. I’ll be cheering for every one of you, and I have no doubt you’ll continue building amazing things.
So…what’s next? First, a bit of a pause.
I’m heading to Canada for part of the summer. I’ll spend some time flying, riding my bike, writing code, and building a few AI projects that have been sitting in my notebook for far too long.
Professionally, I’m launching a consulting and advisory practice focused on applying AI to go-to-market execution and executive operations. Not AI for the sake of AI, but practical systems that help leadership teams sell smarter, operate faster, and make better decisions.
I’m also interested in board and advisory opportunities while exploring
entrepreneurial ventures, building, investing in, and potentially acquiring founder-led businesses where technology and AI can unlock the next phase of growth.
Most of all, I’m looking forward to reconnecting.
If you’re a former colleague, customer, founder, investor, or friend, I’d love to catch up. If you’re thinking about how AI can transform revenue organizations or executive operations, exploring an acquisition, or building something ambitious, I’d genuinely enjoy the conversation. Coffee, a Zoom, or even a flight over the Bay all work.
Transitions create space for new ideas, new partnerships, and new adventures.
I’m excited to see where this next chapter leads.
Venkata Pingali liked this
Venkata Pingali liked this
Huge congratulations to Prof. Hong Wang for winning the Fields Medal, the highest honor in the field of mathematics widely considered as the “Nobel Prize of Mathematics”! Prof. Wang was recognized “for her work in harmonic analysis and geometric measure theory, including applications of multiscale and decoupling techniques to the local smoothing conjecture for the planar wave equation, and major advances in Fourier restriction, Falconer distance sets, Furstenberg sets in the plane, and the Kakeya problem in three dimensions.” Kudos! #fields #medal #mathematics #prize #conjecture #dimensions #congratulations #nobel #honor #recognition
Venkata Pingali liked this
‘Het was intens, inspirerend en ook gewoon plezant. Veel gelachen, veel geleerd. Ik heb het gevoel aan een snellader gehangen te hebben. Dit smaakt naar meer, absoluut.’
Merci dat je erbij was, Christophe 🤙!
Venkata Pingali liked this
Benieuwd wat Christophe Frère van zijn Outspire hike vond? ⚡⚡⚡
‘Het was een fantastische driedaagse. Fijne mensen, een heel complementaire groep en vooral veel inspirerende oefeningen. Wat mij bijblijft: iedereen zoekt. Ongeacht waar je staat als ondernemer. En iedereen heeft een verhaal en daar wordt hier echt naar geluisterd.’
‘De opdrachten verrasten me. Ze halen je uit je comfortzone en brengen dingen naar boven die je niet verwacht. In combinatie met de natuur, het bewegen en de groep zorgt dat voor inzichten en ideeën die je alleen nooit zou krijgen.’
‘Het was intens, inspirerend en ook gewoon plezant. Veel gelachen, veel geleerd. Ik heb het gevoel aan een snellader gehangen te hebben. Dit smaakt naar meer, absoluut.’
Het was geweldig stappen met je te mogen zetten, Christophe! 🙌
—–
Kom mee op Outspire hike en ervaar zelf hoe creativiteit en co-creatie jouw onderneming echt vooruit helpt. Jeff Houwen ⚡️ is creatieve gids, Olivier De Smedt wandelgids.
❗ Schrijf nu in voor 2027! (2026 volzet)
Sardinië 🇮🇹: 8-13 maart / 16-21 maart / 8-13 nov / 16-21 nov
Franse Ardennen 🇫🇷: 16-18 april / 30 april-2 mei / 1-3 okt / 15-17 okt
Finland 🇫🇮: 6-11 sept / 14-19 sept
Sussex 🏴: 28-30 mei (laatste plaats) / 4-6 juni
Waar wacht je nog op?
Venkata Pingali liked this
Venkata Pingali liked this
I am so glad I get to work with Bryce Collins and Erik Johnson. The best teammates make working together fun. Yesterday, we were able to pitch Abasure and our traction for the Sandbox USU Internal Demo Day.
It has been amazing to see the progress I’ve made personally, we’ve made as a team, and the business has made over the course of the last 7 months.
We truely are building something that is making a difference and are very excited for what is ahead. If you know someone who works in autism treatment please connect me!
Experience & Education
Carver Agents
**********
******** ****
**********
** ****** ******** *****
** *********
********** ** ******** **********
****** ** ********** * *** undefined undefined
****** ********* ** *********** ******
******** ** ********** * ***** ******** *******
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
or
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
See who you know in common
Get introduced
Contact Venkata directly
Dermot McGrath 麦德蒙
ZenGen Labs
• 3K followers
I spoke with Supreeth R Koundinya at Analytics India Magazine about how state capital became the default funding mechanism for Chinese AI, and why some founders are choosing it over private VC.
Private VC fundraising in China fell to decade lows by 2024 and to less than a quarter of that level in 2025. A-share IPO approvals went from 400+ in 2022 to roughly 100 in 2024. The exit flywheel that VC depends on stopped turning. Government capital filled the vacuum, reaching 81% of LP commitments into Chinese VC and PE funds by 2025 (Rhodium Group).
The state capital operates through two channels. The investment track puts patient capital (13-15 year horizons) into frontier AI companies, with access to Huawei’s domestic compute allocation and a path to Hong Kong IPOs. The procurement track scaled to $4.2B across 7,500 government AI contracts in 2025, up from $900M across 1,520 in 2024.
Liang Wenfeng is closing DeepSeek’s first external round at ~$7.35B, putting ¥20B of his own money in (40% of the round), retaining ~84% equity and near-total voting control. He stated publicly he would only accept “national-team capital and entity-industry capital” (国家队资本和实体产业资本). State capital aligns with DeepSeek’s technical roadmap (reducing NVIDIA dependence and optimizing for Huawei hardware), avoids US regulatory exposure from American-linked VCs on the cap table, and provides the compute infrastructure the company needs to scale. A founder keeping 84% equity and total voting control while taking state money is not being coerced into anything.
Foreign capital isn’t locked out. The terms are: keep the technology in China, let Chinese investors capture the upside. For many frontier Chinese AI companies right now, state capital is the preferred source of funding.
Full piece:
https://lnkd.in/gj7SGF6Y
2
Aadil Shaikh
V5 Global
• 237 followers
Recommendations for Builders
Constitutional AI for Deployment Now
Implement CAI with 15–20 domain-specific principles
Use chain-of-thought (CoT) in preference evaluations (~75% accuracy vs. 65% baseline)
Deploy human helpfulness labeling only; automate harmlessness evaluation
Quantify harmlessness via Elo scores from crowdworker comparisons
Mechanistic Interpretability as Safety Layer
Train sparse autoencoders on model activations (feasible up to ~70B params)
Identify causal circuits for critical behaviors (e.g., refusals, reasoning)
Use automated interpretability (EleutherAI pipeline) to scale feature discovery
Set monosemanticity threshold; alert on polysemantic features (reward hacking indicator)
Automated Red Teaming Before Deployment
Use tools (Giskard, Maxim, PointGuard) to detect 40+ vulnerability classes
Train attacker agents on diverse goals (prompt injection, jailbreaks, hallucinations, PII leakage)
Set false positive thresholds; monitor trend over time
Map results to OWASP LLM Top 10 & NIST AI RMF for compliance
Containment for Development & Research
Light containment: Virtual machines, firewalls, network isolation (low overhead)
Medium containment: External supervision, instruction-level monitoring (1.5x compute)
Defense-in-depth: Combine OS-level (seccomp), hypervisor (KVM), and monitoring (tripwires)
Sensitive info management: Restrict containment details, personal data, hacking tools
Corrigibility as Practical Safety Principle
Implement separate utility heads (deference, switch-access, truthfulness, low-impact, task)
Use lexicographic combination with strict weight gaps (higher heads dominate)
Deploy human oversight fallback controls (kill switches, credential revocation)
Verify finite-horizon safety; audit in “decidable islands”
Continuous Monitoring & Incident Response
Monitor model drift, anomalies, reward spikes (real-time SIEM + UEBA for AI)
Quarterly penetration testing; forensic logging for compliance
Establish incident response SLAs: detect in <12 hours, remediate in <24 hours
Maintain audit trail for regulatory (ISO 42001, NIST AI RMF, EU AI Act)
Ashish Kakran
Seligman
• 12K followers
We at SIERRA Ventures are thrilled to lead the $8M Seed round in smallest.ai! We feel honored to partner with Sudarshan Kamath and Akshat Mandloi as they build an important company in the Voice AI space.
We’ve now entered a new era. Voice AI agents are no longer research projects or demos – they’re deployed in production. Whether it’s buying or selling a house, scheduling a doctor’s appointment, or returning a purchase, there’s a growing chance you’ve already interacted with an AI voice agent without realizing it.
Why Smallest AI:
💡 Team: Sudarshan and Akshat are customer-obsessed technical founders who understand what enterprises truly need.. agents that are fast, secure, and production-ready.
💡 Industry leading performance: The company has some of the fastest, highest-quality voice AI models in the world (Lightning Text to Speech) that outperform Elevenlabs, Cartesia across all parameters. Their SLM Electron outperforms GPT 4.1 Mini in both quality and latency for real-time conversational use cases.
💡 Integrated Voice stack: Delivering a natural voice interaction requires a complex workflow under the hood integrating components like ASR, models, TTS. By optimizing every step in the loop, from low-latency TTS to contextually coherent responses, they are helping to deliver real-time, human-like conversational AI that sets a new benchmark for performance.
💡 Small models, big impact: Nvidia recently published a research paper – “Small models are the future of agentic AI”. Smallest AI builds the high-accuracy models that customers can run at the edge consuming little power.
💡 Happy customers: customers consistently told us how much they enjoyed working with the team.
For anyone building enterprise voice AI agents, please reach out and we’ll be happy to connect you with the team!
#voiceAI #ai #tts #asr #slm
424
32 Comments
Pranay Prateek
SigNoz
• 31K followers
🚀 Our article on issues with LLM observability currently and the lack of standards there is in top 10 in HN.
Inspired by my conversation with Pranav Raj S, it seems lots of folks are facing such issues
We discuss how building, debugging, and improving AI agents in production gets messy fast. There’s multiple competing standards for default libraries for LLM observability. And many such libraries like OpenInference which claim to be based on OpenTelemetry don’t strictly adhere to it’s conventions. This introduces problems for users who are trying to get better observability across their stack.
Complete article in comments
41
2 Comments
Find curated posts and insights for relevant topics all in one place.
Cloud-Based AI Solution Design Patterns
AI Coding Agents with GitHub Copilot and Cursor
Creating Agents and Custom Copilots in SharePoint (No Code Required)
Venkata can introduce you to 5 people at Carver Agents
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
This data page is restricted to registered members. Please sign in to unlock full access to this intelligence table.
Sign In to Unlock Access
Welcome Back 👋
Sign in to access your account and continue where you left off.
Welcome aboard 🚀
Create your account to access everything you need.
newsletter subscription
Do you have a specific requirement to engage with a particular community? Then we’ll be glad to devise a custom program just for you. Please fill the form below and we’ll contact you as quickly as possible.
newsletter subscription
Get a curated mix of deep-dive research and real-time market signals.
See what’s next. Act with confidence. CMR combines market intelligence, strategic advisory, original research, and industry engagement to help organizations navigate change and accelerate growth.
Fill out the form below and we’ll connect our analyst
Access real-time industry intelligence, expert analysis and strategic insights to stay ahead of market shifts and competitive dynamics.
Expert guidance to help you navigate complex business challenges and accelerate your growth trajectory.
Leverage our deep industry knowledge to create high-impact marketing campaigns that resonate with your target audience.
Venkata Pingali replied to a comment
3w