Siddhartha Singh

How Generative AI is reshaping data architectures, from traditional pipelines to prompt-centric, semantically aware systems. This post covers prompt orchestration, embedding pipelines, and real-time inference with observability and compliance in mind. https://lnkd.in/gTWxsDBU

25

5 Comments

"For these models to reach their peak value, you need to train them not just on publicly available data, but make privately owned data available for those models as well," - Ellison https://lnkd.in/gM6FskzJ

4

Low-Level Design translates architecture into implementation using patterns, structure, and conventions. As generative AI accelerates code generation, the real value lies in knowing when human design insight is needed to complement and refine AI-generated outputs.

16

2 Comments

Apache Spark’s new "Real-time" Mode in Structured Streaming delivers millisecond-level latency for ultra-fast, continuous stream processing. It will be interesting to see how it will perform compared to Flink!

7

Proud to see our work with Nimbus Therapeutics featured in the AWS re:Invent keynote! Our team developed drug-discovery aware LLMs using Nova Forge, combining Supervised Fine-Tuning (SFT) and Reinforcement Fine-Tuning (RFT) to create custom model tailored for Nimbus. Thanks Andy Lapastora Krishnateja Killamsetty Elyse (Jiali) Zhang and Clement PERROT for the exceptional work. Also thanks to the wonderful AGI and Nimbus team for making this collaboration a success. cc: Hannah Marlowe, PhD Sri Elaprolu Taimur Rashid Sharlina Keshava Caleb Wilkinson Alexandra Fedorova

45

1 Comment

Couldn’t agree more. In any technology spectrum, there are always two extremes—and the real value for most people often lies in finding that optimal midpoint. In AI, that’s where the benefits of decentralized intelligence truly shine.

5

Had a great time attending the Product Thinking and the Future of EPD(Engineering, product and Design) as an Engineering Leader event hosted by ELC today! A big thank you to ELC and Commure for organizing this thoughtful and engaging session—and to the amazing panel, including Dhruv Parthasarathy, Nissim Lehyani, James Tyack,Patrick Gallagher, for sharing their candid insights on how the traditional boundaries between Engineering, Product, and Design are rapidly evolving in the age of AI. Here are a few takeaways that really stuck with me: 🔹 Companies are shifting away from static PRDs and increasingly using prototypes as a discovery tool—validating ideas with customers early and often. 🔹 Hiring is changing: it’s less about checking boxes and more about finding problem solvers and clear thinkers who can adapt and lead. 🔹 When it comes to LLMs, everyone has their own “AHA moment.” Until then, keep exploring. You’ve essentially got the smartest person in your pocket—start using it to solve real problems. Excited to bring these insights back to my team and continue experimenting at the intersection of product, engineering, and AI! #EngineeringLeadership #AI #ProductThinking #ELC #LLM #TechLeadership #Prototyping #EPD

9

2 Comments

We shipped automatic snapshot expiration for R2 Data Catalog today. With this new feature you can define retention policies (age thresholds and minimum snapshots to keep) so old snapshots are cleaned up automatically, reducing metadata overhead and storage costs while keeping table performance healthy. https://lnkd.in/eAD5rKd8

46

Eval is not unit testing. If you treat evals like static pass/fail scripts, you’re missing their real purpose. Think of evals as a PRD for your AI agent’s behavior: a living, evolving contract for what “good” means in production. And like any PRD, it should have observability built in so you can track outcomes, pinpoint weaknesses, and refine in an automated, continuous loop. Improvement without measurement is limited and imprecise. Evals give you the clarity you need to systematically improve your app, not limited at the feature level, but across the full agent lifecycle. They tell you where you’re drifting from the product’s intent, why performance is slipping, and how to course-correct with precision. Instead of snapshot testing, you’re running a feedback engine that constantly learns from both successes and failures. In the application layer, it’s about setting big-picture goals, like “help a customer dispute a fraudulent transaction” or “recommend the right investment product for their profile.” Then you throw everything at it half-filled forms, conflicting details, urgent time limits and see if it still gets the job done. The point isn’t to check if it spits out a canned line, it’s to see if it can solve the problem accurately, fast, and in a way that stays within compliance and cost boundaries. My recommendation approaches to make evals work for AI agents: 1. Define capabilities, not scripts – Focus on what success looks like rather than exact matches. 2. Simulate real-world variability – Feed in unpredictable, noisy, and edge-case inputs. 3. Instrument for observability – Capture reasoning traces, model calls, and decision trees for explainability. 4. Automate refinement loops – Turn eval failures into direct triggers for retraining, prompt tuning, or guardrail updates. 5. Score across multiple dimensions – Measure accuracy, latency, cost efficiency, compliance, and user experience impact in parallel. When done right, evals validate your AI agents and steer them. They transform from a passive QA step into the heartbeat of your improvement cycle, ensuring every release moves the agent closer to its ideal, measurable definition of “good.”

40

3 Comments