About
I am an AI Architect and Lead Data Scientist focused on building production-grade agentic…
Services
Activity
14K followers
-
Prasanth Kumar Kolluru shared thisExcited to present at Data + AI Summit 2026 with the Cummins team — sharing how we built and deployed our first AI Agentic System for supply chain. What worked, what didn't, and what we wish we knew earlier. The honest version, not the highlight reel. #DataAISummit #innovation
-
Prasanth Kumar Kolluru shared thisAgentic Brain We're entering a new era of agents — where a single agentic brain can handle every part of a business, end-to-end. To test this, I built a prototype: an AI agent running a fully autonomous coffee shop with a $100K budget and one mission — generate $700/week in net profit. Over 48 days, the agent did it all — it identified job roles, posted openings on LinkedIn through an MCP server, filtered candidates, and scheduled interviews. It prepared government license documents so accurately it genuinely surprised me. It managed store operations (shifts, schedules, customer flow) and inventory (auto-reordering supplies, predicting demand, avoiding waste). It tracked every transaction, wrote daily journals to reflect on its performance, and even ran its own experiments — like extending store hours on specific days, which delivered a $157 revenue lift. Anything above $500 comes to me for approval, but everything else? It runs on autopilot. Github Repo - https://lnkd.in/ePQZgGm3 This is just the beginning — and I'm planning to run it for a full year next. 👇 Watch the full demo below to see how it works in action.
-
Prasanth Kumar Kolluru posted thisWeek 2 Problem: E-commerce company losing $2.1M annually with quarterly pricing updates while competitors changed prices every week. The Solution: Built a swarm of autonomous AI agents that work 24/7 to optimize pricing in real-time. Business Impact • +$2.1M annual revenue (18% increase) • 90% less manual work for pricing teams Week 2: How 5 Autonomous Agents Generated $2.1M Extra Revenue The 5 Agents That Transformed Pricing 1. Product Catalog Agent (150K items) Business Impact: $420K annually Monitors product costs and current prices 24/7. Specific Example: When supplier costs for iPhone screen protectors dropped 12% due to bulk purchasing, this agent automatically flagged 2,500 related products within 2 hours. Result: $89K immediate margin recovery on a single product category. 2. Transaction Agent (89M transactions/year) Business Impact: $630K annually Tracks real-time sales patterns to identify high-demand periods. Specific Example: A $25 phone charging cable went viral on TikTok, sales jumped from 50 units/day to 2,500 units/day overnight. This agent detected the 50x demand spike within 30 minutes and recommended a 20% price increase while demand was at peak irrationality. Result: Generated $47K additional revenue in first 48 hours of viral trend. 3. Competition Agent (2.3M updates/day) Business Impact: $315K annually Monitors 127 competitor websites every 15 minutes. Specific Example: Best Buy increased prices on Sony 65" TVs by 18% during supply chain shortages. This agent detected the change within 15 minutes and recommended matching the price increase. Since consumer demand remained strong despite higher prices, the system captured additional margin without losing sales volume. Result: $112K additional revenue on TV category in first week of shortage. 4. Inventory Agent (4.2M updates/day) Business Impact: $420K annually Watches stock levels across 12 warehouses. Specific Example: Hurricane approaching Florida threatened supply chain for portable generators. This agent noticed inventory dropping faster than expected and proactively increased prices by 25% on remaining stock in East Coast warehouses, while demand was highest due to storm preparation. Result: $67K additional revenue from limited inventory before supply chain disruption. 5. Customer Behavior Agent (560M records) Business Impact: $315K annually Analyzes how different customer segments respond to prices. Specific Example: Discovered that customers who browsed for >5 minutes and added items to cart but didn't purchase were 3x more likely to buy with 15% discount vs. 25% discount (which made them question product quality). This agent automatically served personalized offers, increasing conversion from 8% to 22%. Next week: How autonomous supply chain agents cut inventory costs 25% while hitting 99.2% on-time delivery. #AI #AgenticAI #Ecommerce #RevenueOptimization #BusinessImpact
-
Prasanth Kumar Kolluru posted thisStarting this week, I’ll be sharing one real-world AI/ML use case every week that solves high-impact business problems — from churn prediction to fraud detection, demand forecasting, and more. 💬 Have a business challenge you’d like me to cover? Drop it in the comments — and I’ll break it down in a future post! 🚀 Building an Intelligent Churn Prediction System with Autonomous Agents 💡 The Challenge Customer churn was silently draining $2.3M annually from a leading SaaS company. With a 15% monthly churn rate, the team needed more than just dashboards: ✅ Real-time churn prediction ✅ Automated, explainable retention insights ✅ Actionable strategies—not just alerts 🧠 The AI-Powered Solution: Autonomous Agent Architecture We designed a multi-agent intelligence system that does the heavy lifting: 🧩 Data Agent: Cleans & validates 2.1B+ records across 5 sources 🛠️ Feature Agent: Builds 47 dynamic features from usage, billing, support, and product signals 🧮 Model Agent: Uses XGBoost + RF + Neural Net for 91% AUC 🎯 Action Agent: Prescribes targeted, personalized retention campaigns 📊 What We Engineered 1.2M customers analyzed across: Usage patterns 📉 Billing behavior 💳 Support trends 🛟 Product engagement 🚀 Key features like: usage_decline_30d payment_delinquency_score feature_adoption_velocity ⚙️ Deployment 🧵 Real-time scoring every 4 hours 🐳 Deployed via Kubernetes + Flask APIs 📉 False positives reduced by 60% 📈 89% of churners flagged 2 weeks in advance 💰 Business Impact 💸 $736K saved annually ⏱️ 70% reduction in manual review time 📈 Lift = 4.2x over rule-based systems 🔍 Precision@10 = 0.84 🎯 Why It Matters This isn’t just churn prediction—it’s autonomous decision intelligence in action. Your customers won’t tell you when they’re about to leave. But your system can. Let’s build the future of customer success—proactive, not reactive. 🧵 Want a deeper breakdown of the architecture or code? Comment "CHURN" and I’ll share the technical walkthrough 👇 #AI #ML #CustomerSuccess #ChurnPrediction #LLMs #XGBoost #NeuralNetworks #AutonomousAgents #DataScience #Python #SaaS #Retention #Flask #Kubernetes #EnterpriseAI #GenAI #AgenticAI #Learning
-
Prasanth Kumar Kolluru posted thisWhen Complexity Meets Cost: A Common Challenge in Manufacturing A major manufacturer is struggling to manage cost optimization across customized engines for agriculture, mining, defense, and automotive sectors. Each engine has 300+ parts—a small cost increase in one part can ripple across multiple engines, customers, and margins , creating a complex web of impacts. Here’s the reality: 🔴 Challenges : Tracking which parts have gone up and why Finding real cost savings across thousands of components Understanding margin impact—especially on fixed-price contracts 🕒 Manual analysis takes too long. 📊 Dashboards are fragmented. 💡 Insights are buried under data. So I am trying to understand how teams are solving this type of complexity? Are there tools or strategies that actually help?
-
Prasanth Kumar Kolluru shared thisHello Everyone, I recently watched a fascinating show about the Tata Group and discovered some key insights from their 156 years of innovation and business success. I believe these takeaways are valuable for any entrepreneur or business: A Larger Purpose: The Tata Group has always been driven by a mission to contribute positively to the nation. Every entrepreneur and business should have a broader purpose beyond profit, aiming to make a meaningful impact on society. A Pioneering Mindset: For instance, Jamshedji Tata’s determination to establish India’s first steel plant in the early 1900s is a testament to innovation and resilience. He faced numerous challenges but persevered by traveling to the US to bring together the right people and resources. Commitment to Quality and Principles: The Tata Group has consistently focused on producing high-quality products while adhering to strong ethical principles. This dedication to excellence and integrity ensures that their offerings meet the needs and expectations of people. I hope these insights inspire you as much as they inspired me. Whether through small or large contributions, every effort counts towards creating a positive impact on our nation. share you opnion in the comments This book tells you many stories about the tata group how it has developed and vision to different business https://lnkd.in/gFuuSRskJamsetji Tata: Powerful learnings for corporate successJamsetji Tata: Powerful learnings for corporate success
-
Prasanth Kumar Kolluru shared thisPrasanth Kumar Kolluru shared thisThe latest AI News is out, If you enjoyed this issue, please subscribe to this newsletter. Thanks to supporters and 1 Million+ readers🙏 Also, I partnered with Nvidia to give away one Geforce 4080🔥 to one lucky person and 10 vouchers to those who attend this free event. #innovation #artificialintelligence #technology #gtc23 #intelxeonChatGPT and Generative AI Tips & News; Google says Bard isn’t just about search; Elon Musk says that humanoid robots will eventually outnumber people.ChatGPT and Generative AI Tips & News; Google says Bard isn’t just about search; Elon Musk says that humanoid robots will eventually outnumber people.
-
Prasanth Kumar Kolluru shared thisPrasanth Kumar Kolluru shared this𝙎𝘾𝙃𝙊𝙊𝙇 𝙈𝙀𝙈𝙊𝙍𝙄𝙀𝙎 ❤️❤️ I came across these wonderful school memories by Paperboat that we all shared during our school years. Those were incredible times.❤️ Follow Geeta Rautela for more. #schooldays #memories
Prasanth Kumar Kolluru replied to a comment
3mo
Senthilvel Rajan Kumaresan thank you 😊
-
Prasanth Kumar Kolluru reacted on thisPrasanth Kumar Kolluru reacted on thisToday, Uber promoted me to customer. It's not the career move I had planned, but here we are. I was impacted by today’s organization-wide layoffs, along with many colleagues across Uber. It came as a shock to me and I'm sure to many others on my team/org. Navigating re-orgs and market-shifts is part of the professional journey, and honestly I don’t take these decisions personally. Reorgs happen and priorities change. I’m still proud of the work I did and the people I worked with and I can't think of a better place to start as a New Grad. I learned a ton from my team at Uber and saw firsthand what a world-class competitive intelligence team looks like. My day-to-day work focused on generating new insights and making the supporting data pipelines more reliable through better monitoring, alerting, and data-quality checks. I’m especially proud of winning the Toronto site award in the 2026 Uber Dispatch Challenge, where I built a model-agnostic AI tool-chain & harness that automated log analysis, testing, deployments, and ticket-to-review workflows. I’m now reaching out to my network and looking for leads on new grad/junior software engineering opportunities. I’ve attached my resume, and I’d appreciate any introductions or roles you think I could be a good fit! I’m excited to bring the same drive and enthusiasm wherever I land next. To everyone else affected today: I’m rooting for you. EDIT: I found a new role! https://lnkd.in/p/gSRDepeH #OpenToWork #SoftwareEngineering #Toronto #Tech
-
Prasanth Kumar Kolluru liked thisPrasanth Kumar Kolluru liked thisTake a Different Approach - When Less Produces Far More In Business Here is a young woman who entered the Shark Tank with a clear idea, a confident presence, and a strategy that defied the norm. She wasn’t there to sell 24/7. She came with a different mindset, which was an understanding of how to position value, rather than chase customers each day. What happened next? One product drop achieved $260,000 in 24 hours. She didn’t flood the market. She didn’t use sales or discounts. She offers a single monthly opportunity. This is the essence of thinking differently. While others scramble to stay visible, she builds her brand between launches. While many chase validation through constant activity, she builds desire through absence. In my role, I look for what sets people apart. In this case going from purely performing (sales) to positioning. From chasing to creating customers. You build it by stepping into strategy that’s as bespoke as the life you’re creating. Leadership today demands more than pure ambition. It asks for discernment. You can’t build influence doing what everyone else is doing. You don't have to follow the crowd. Think differently, execute smartly, and watch the results that you will achieve. Are you a leader, entrepreneur or someone wanting to make a change? What would it look like to shift from purely working hard to greater results from working wisely? Repost if you feel this would help others.
-
Prasanth Kumar Kolluru liked thisPrasanth Kumar Kolluru liked thisMy father always used to say… You don’t really know your boss until you mess up big time… And honestly, he was right. The true measure of leadership isn’t seen when goals are met. It’s revealed when mistakes are made. A leader has two choices in those moments: - React with anger, and build fear. - Respond with calm, and build trust. That single choice changed everything. One shuts people down. The other opens people up. One creates silence. The other creates learning. Because.. This is what many leaders miss: People don’t remember the size of the mistake. They remember how their leader made them feel in that moment. And The greatest leaders are not those who prevent every error, but those who transform errors into opportunities, and failures into foundations. And that’s exactly what this video captured so beautifully…🌱 A boss choosing; ✅ Patience over punishment. ✅ Guidance over blame. ✅ Trust over tension. A reminder every leader should carry. 💟 ——————————————————————————— Video Source: Instagram
Experience & Education
Licenses & Certifications
Projects
-
Search pattern prediction
- Present
An insurance company has a lot of policies on their list. But when the user types for a keyword it should be able to suggest some policies based on the previous searches and also mostly purchased polices.
-
Dynamic routing
-
Dynamic Routing is a business intelligence project which will help cases to route the automatically(Bank loan case, Insurance application case). For example, the user creates a request to apply for a bank loan so here the system is intelligent enough that decide whether he is good or bad based on his previous data. If he is bad what are additional checks required and where should be it routed. In this, it helps cases to route to different operators and work queues.
-
Self-driving Car
-
The project is used to estimate the steering angle based on the past data of the steering angle and its front view camera images.
-
Music generation using LSTM
-
The project is used to design new music based on the existing musical notes.
-
Loan approvals estimation
-
A bank to predict whether to issue a credit card or a loan to a particular person based on his cibil score, salary, and a few more things.
-
Intelligent sport allocation using image recognization
-
we have implemented this project internally in our company. So we have some 3 TT tables and some billiards tables in out office and we are around 1500 people so a lot of people of interested to play but they go to the play zone and see someone playing and fell bad and come back. So we have developed an intelligent system so that it would give suggestions to the users when the TT tables are free to play so that maximum users can play the game
Other similar profiles
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Prasanth Kumar Kolluru
1 other named Prasanth Kumar Kolluru is on LinkedIn
See others named Prasanth Kumar Kolluru