Our paper "Gradient-Based Join Ordering" (with @maribelacosta ) will be presented at IJCAI 2026! Join ordering, deciding the order in which to evaluate the joins of a database query, is a classic and discrete NP-hard problem, and one of the most important in any database engine. Current engines approach it using discrete algorithms like dynamic programming, greedy heuristics, or genetic algorithms. In our new paper, we explored a different option. If the cost model is differentiable (e.g., a neural network), why not search using a gradient? For that, we relax discrete query plans into a soft adjacency matrix, i.e., a continuous superposition of plans, and descend the cost landscape directly via gradient descent. Differentiable structural penalties (degree constraints, acyclicity, left-linearity) combined with temperature annealing ensure the optimization converges to a valid, discrete plan. Interestingly, although the cost model is only ever trained on valid, discrete plans, the relaxed space between them turns out to be smooth and informative — the local gradient reliably points towards better plans. Key results on two standard benchmarks: • Plan quality matches and often surpasses discrete baselines • Only ~10 cost model evaluations are needed, vs. hundreds for randomized search • Runtime of 40–200 ms with more favorable scaling in query size than greedy search We think this is a first step towards query optimizers that treat plan search as continuous optimization rather than combinatorial enumeration. #IJCAI2026 #QueryOptimization #DatabaseSystems #GraphNeuralNetworks #KnowledgeGraphs

📢 Join us for the Pre-SIGMOD Workshop in Bangalore! 📅 30 May 2026 📍 BMS College of Engineering Featuring talks by Carsten Binnig and Pavan. Open to students, faculty, researchers, and industry professionals. 📝 Register: acmblrindia.hosting.acm.org #SIGMOD #DatabaseSystems

📢 Tutorial Session at #SIGMOD2026 Data Agents: Levels, State of the Art, and Open Problems Join renowned experts Yuyu Luo, Guoliang Li, Ju Fan, and Nan Tang as they explore the evolving landscape of data agents, current advancements, and key open challenges shaping the future of data systems. 📍 Sheraton Grand Bengaluru 📅 Wednesday June 3, 2026 ( 1:30 PM - 5 PM) #SIGMODPODS2026 #SIGMOD #PODS #TutorialSession #DataAgents #AIResearch #DatabaseSystems #ResearchCommunity #Bengaluru

We are proud to welcome Bosch Global Software Technologies as a Platinum Sponsor for SIGMOD PODS 2026 in Bengaluru, India! Their support strengthens a global platform dedicated to advancing innovation, research, and collaboration in the database and data management community. 📅 May 31 – June 5, 2026 📍 Sheraton Grand Bengaluru Looking forward to an inspiring and impactful conference experience together. #SIGMOD2026 #PODS2026 #Bengaluru #GlobalConference #DatabaseSystems #ResearchCommunity #TechConference #Innovation #DataManagement

Data is only as powerful as the trust behind it. 📊🔍 Join this insightful tutorial session as leading experts Chrysanthi Kosyfaki and Paul Groth explore the evolving landscape of data provenance from foundational concepts to emerging challenges and future opportunities. 📅 June 2, 2026, 11 AM - 12:30 PM 📍 Sheraton Grand Bengaluru #SIGMOD2026 #PODS2026 #DataProvenance #DataScience #ResearchCommunity #DatabaseSystems #AcademicConference

Read the article "Buy vs. Build for Time-Critical Sensor Data Fusion" in @ElectronicDesgn t.ly/ICsvU “Real-time” definitions & insight into time-critical #sensordatafusion and its role in #databasesystems. #EmbeddedSystems #timecritical #RealTimeData #DBMS

Good day! Hero tcRfiWtYG @databasesystems @metracoin @Ella79685593 @hao833827034739 @jakecasillas @Rayon70173888 @r_aguenza @a7x_hendri @dcDKHeSijbwpfzV @mac2201 @Melissa_brell @chiakarhee1 @ArttaSahu @JamesZhangHOHO @s_stronz @soumendey314 @Bhavya136 @moha_noe @hanxiao1300 @basusahani07 @anees6942 @InessaMari @adylon7 @salas5657 @Gaming37203176 @rM8SEHMkTK6758 @kacheng @iris_ETHH @TaglioliGiada @xiverifund @wilson_gloryhk @JonathanL64 @ruihayun @SouravM41873434 @yamadamasao0 @IGI_official_ @sharuchuru @FiCraze @S_Hannestad @dkavtaradze1994 @nileshsshah @crypt0trader5 @praise8w @sunyfish9527 @Imrankhan510964 @SonicSimulation @SumadhurTej @shehbaz4u @louisdazy @Gme4L853 #SpaceX SIEAuHUbT

Automating Database-Native Function Code Synthesis with LLMs 🎯 Writing database functions is still painfully manual What if LLMs could handle it end-to-end—correctly? 📌 Problem Database “native functions” (like date_trunc) are complex: multiple internal units, files, and dependencies. Current LLM coding tools miss key details → wrong references, missing pieces, broken outputs. 💡 What this paper tries Build DBCooker: an LLM system specialized for database function generation. Instead of raw code gen, it understands structure first, then fills in the logic. 🧪 How it works Breaks a function into units dependencies using static analysis (like mapping all moving parts first). Generates a pseudo-plan (a structured blueprint) before writing code. Uses fill-in-the-blank coding 3-step validation (syntax → rules → actual outputs). 🚀 Why it matters 34.55% higher accuracy vs existing methods across SQLite, PostgreSQL, DuckDB. Can even create new database functions that don’t exist yet. 🤔 What’s next Still depends on correct planning—bad plans can cascade into errors. What happens when this scales to entire database systems or migrations? YouTube: youtu.be/Cns2Ql-dXQU 🔗 arxiv.org/abs/2604.06231 🏷️#AI #MachineLearning #GenAI #LLMs #DatabaseSystems #CodeGenerati

#FrontPageLIVE 🔴 @Wipro just quietly signalled the end of the headcount-led IT model, Jensen Huang is pushing back hard on anyone calling @NVIDIAAI chips a commodity, and someone needs to answer a very uncomfortable question about who is actually paying for India's AI e-waste. We also have Professor Sudarshan breaking down how databases have evolved from relational roots to AI-driven systems, and why that history matters more than ever right now. We're going live on AIM Front Page. See you there: youtube.com/live/daoTO4Yj3QE #FrontPageLIVE #AIMNetwork #Wipro #NVIDIA #JensenHuang #AINews #IndiaAI #TechNews #DatabaseSystems #AIInfrastructure

6/ The deeper I go into this space, the clearer it becomes: Some of the most interesting systems problems today live at the intersection of compilers and data infrastructure. And it’s a fascinating place to build. #DatabaseSystems #CompilerEngineering #SystemsProgramming

📊🤖 Highly Cited Paper (2024) in MAKE GAN-Based Tabular Data Generator for Constructing Synopsis in Approximate Query Processing: Challenges and Solutions 📖 Read the full article: mdpi.com/2504-4990/6/1/10 #GAN #ApproximateQueryProcessing #DatabaseSystems

Learn DBMS vs RDBMS differences with examples, pros, cons, and use cases to choose the right database system for modern applications. bit.ly/DBMSvsRDBMS #DBMSVsRDBMS #DatabaseSystems #DatabaseManagement #DatabaseComparison #SmartDatabaseSelection

Prof. Ahana Pradhan, Department of CSE, #IIITB, was invited to present her paper titled “Defog: Extending Hidden Query Extraction to Group Filters”, co-authored with Prof. Jayant Haritsa, IISc, at the India AI Impact Summit in Delhi recently. The paper was accepted at CODS 2025 and received the Best Paper Runner-Up Award at the conference. At the AI Summit, it was showcased under the Global South Research Showcase category, which highlighted papers from top-ranked AI conferences over the past two years authored by researchers from South Asian countries. . . #IIITB #IIITBangalore #CODS2025 #IndiaAIImpactSummit #ArtificialIntelligence #AIResearch #DatabaseSystems #DataScience #ComputerScience #ResearchExcellence #WomenInTech #WomenInAI #GlobalSouthResearch #AcademicResearch #ConferenceHighlights #BestPaperAward #CSE #DelhiEvents

🏎️ Day 17 of The Product-Engineer-Max Challenge Completed the full SQL for Data Science roadmap — from SELECT & WHERE to JOINs, GROUP BY, CASE logic, and real interview-level problems. Check it out: codingismeditation.github.io… #ProductEngineer #SQL #DataEngineering #DatabaseSystems

🏎️ Day 16 of The Product-Engineer-Max Challenge Completed the full DBMS from ACID & normalization to keys, locks, and 3-tier architecture. Check it out: codingismeditation.github.io… #ProductEngineer #DBMS #DatabaseSystems #BackendEngineering #LearnBuildShip

#DocDrop #CloudInfrastructure #SecureStorage #DataCenter #EnterpriseStorage #DatabaseSystems #EncryptedData #CloudSecurity #TechInfrastructure #ScalableSystems #DigitalTrust #CloudComputing #MadeInIndia

𝐀𝐩𝐩𝐥𝐢𝐞𝐝 𝐀𝐈 𝐢𝐧 𝐃𝐚𝐭𝐚𝐛𝐚𝐬𝐞 𝐒𝐲𝐬𝐭𝐞𝐦𝐬, 𝐏𝐌𝐓𝐒/𝐀𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭, 𝐒𝐚𝐥𝐞𝐬𝐟𝐨𝐫𝐜𝐞 🚀 Salesforce is hiring an Applied AI in Database Systems, PMTS/Architect in Hyderabad. Build AI-powered, self-healing cloud databases using ML and LLMs at massive scale. 🤖💾 #SalesforceCareers #AppliedAI #DatabaseSystems #MachineLearning #analyticsinsight #analyticsinsightmagazine Read More 👇 zurl.co/ArwoO

Businesses are replacing traditional single-purpose databases with advanced task-specific #DatabaseSystems like: 📌#DistributedKeyValueStores 📌NewSQL 📌Graph databases

Tutorial Series: Deploying an EloqKV Cluster on EloqCloudIn this video, we walk through how to create your first EloqKV cluster on EloqCloud. Learn how to deploy, connect, and start using a Redis-compatible key-value database with built-in scalability, tiered storage and high performance. Try EloqKV Free: cloud.eloqdata.com Contact us: eloqdata.com/contact#EloqKV #EloqCloud #EloqData #CloudComputing #DatabaseSystems #KeyValueStore #RedisCompatible #ScalableStorage #HighPerformanceComputing #CloudDatabase #DataInfrastructure #DevTools #Developers #TechTutorial #HowToDeploy

Day 62 of learning to code 🧠💻 Today was all about DBMS — transactions, keys, normalization, deadlocks… and my brain is definitely in 2NF, not 3NF yet 😭 Exam mode activated ✅📚 #Day62 #100DaysOfCode #DBMS #DatabaseSystems #StudentLife #ExamPrep #CodingJourney #TechTwitter