Fifthel Winter Edition on 4 & 5 December 2026. Speak, showcase, teach - has.gy/KaLO Attend - has.gy/-fQi

Bangalore, India
“My manager made me come here, and I loved it!!” - Winter 2025 participant Send this to your manager 🙂 Fifthel Winter: 4 & 5 Dec 2026 Book your seat: hasgeek.com/fifthelephant/wi… Submit your work by 11 Oct: hasgeek.com/fifthelephant/fi…
1
140
What makes a data copilot useful to a frontline team? At @Swiggy one practical step was generating the daily reports people had to prepare for their managers. Amaresh Marripudi @marripudiamar explains how understanding that routine shaped the product. ↓
1
1
2
229
Watch the full talk for Swiggy’s journey from text-to-SQL to a data copilot: the product decisions, adoption challenges, and lessons from working with different teams. piped.video/YNT0SoxEAJI
1
2
74
Have a similar case study? The Fifth Elephant Winter Edition 2026 focuses on practical use cases and data analysis. Show us the problem, what you tried, and the evidence of what changed. Share your work on 4 & 5 December. Submit a proposal: has.gy/KaLO
2
38
The Fifth Elephant retweeted
Replying to @GaryMarcus
Been engaged in Ai and also been pointing to mistakes in current AI in my own way wave. I think the FOSS community need to hear from you. If interested, I can explore how to make it happen along with @fifthel or @jackerhack
1
1
96
1/6 One team moved away from ClickHouse. Another used it to speed up proximity search. Two talks from The Fifth Elephant 2026 Annual Conference explore what drove those decisions: different workloads, bottlenecks, and ways to avoid unnecessary work. @oodleai @SahajSoftware
1
1
3
274
6/6 Watch them together: what should you precompute? What work can you skip? Where should it run? The workloads and benchmarks differ. The value is in following each team’s reasoning. What evidence has made you change a database, or how you use it?
1
15
1/3 At IndiaFOSS this weekend? Meet @jaidevd - The Fifth Elephant community member and reviewer. His talk: “Being Bayesian: stress-testing India’s loudest stories with FOSS & open data”. 26–27 Sept, Bengaluru. fossunited.org/indiafoss/202…
1
3
2
246
2/3 What does a 90-hour workweek cost people? What can labour surveys tell us about gig work? @jaidevd explores how to test claims with public data and FOSS, publish analysis others can check - and recognise when the data cannot answer. @IndiaFOSS
1
1
1
138
3/3 Investigated a claim with data? Found something surprising? Bring your question, methods and findings to Fifthel Winter, 4–5 Dec, Bangalore. This is the kind of applied data work we’re looking for. Submit talks/showcases by 11 Oct: hasgeek.com/fifthelephant/fi…
1
1
112
“It was my first 5THEL!! It was fun!!” - Winter 2025 edition participant Make this winter your first Fifthel - 4 & 5 Dec 2026. Get your seat: hasgeek.com/fifthelephant/wi… Submit your ideas by 11 Oct: hasgeek.com/fifthelephant/fi…
1
2
179
AI agents need to remember what they did. Where should that state live? At The Fifth Elephant 2026 Annual Conference, @sunbains made the case for strong consistency. Watch: piped.video/9c7WBxKGbWg @PingCAP
1
2
3
349
The Fifth Elephant retweeted
😙😙 check it out
1/6 What does a good data Showcase look like? At The Fifth Elephant Winter 2025, Aryan Srivastava presented work on using LLMs to extract messy tables from India’s 1951 census. @ary_sriv Old records, a practical problem, and an inventive way to make the data usable.
1
2
109
1/6 What does a good data Showcase look like? At The Fifth Elephant Winter 2025, Aryan Srivastava presented work on using LLMs to extract messy tables from India’s 1951 census. @ary_sriv Old records, a practical problem, and an inventive way to make the data usable.
1
1
3
360
5/6 Need inspiration? Watch Aryan’s talk, “Decoding 1951 census: LLMs extract messy tables”: piped.video/shtbAcdvca8 Think about your own work: what could you show that would make another data practitioner want to investigate, build, or try something?
1
1
43
6/6 Share your work at The Fifth Elephant Winter 2026 Showcase. Tell us what you solved, why you chose your approach, what you learned, and what you can show. Bring concrete examples, anonymized where needed. Submit a proposal: has.gy/KaLO
25