Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

COMPANY · ENTITY #5143

QCon AI

Related event timeline, sources and context from the news index.

EVENT TIMELINE

18

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Niko Matsakis demos Rust Foundation’s Symposium and Agent Client Protocol at QCon AI

Nicholas (Niko) Matsakis argued that Rust pairs well with AI and presented Symposium, an open-source agent-oriented toolchain from the Rust Foundation, demonstrating the Agent Client Protocol (ACP) as a way to standardize agent-to-tool extensibility across developer workflows. He also emphasized how Rust’s type system and compiler guardrails can help agents produce more correct code and enable safer, production-ready systems.

6.0

COMPANIES · 1 SOURCE · InfoQ AI, ML & Data Engineering

LinkedIn engineer details production LLM ranking optimizations including 7B→400M distillation

At QCon AI (covered on InfoQ), Sundara Raman Ramachandran (LinkedIn’s LLM inference lead) described how LinkedIn productionizes LLM-based ranking under tight P99 latency SLAs by distilling a 7B model to ~400M parameters, compressing job contexts into single-token embeddings to boost throughput ~10×, and removing system bottlenecks (custom flash-attention masks, Python GIL workarounds, gc.freeze, non-blocking CUDA memory copies).

7.0

RESEARCH · 1 SOURCE · InfoQ AI, ML & Data Engineering

Dataiku demos Kiji Inspector: sparse-autoencoder explainability for agent tool selection

At QCon AI, Hannes Hapke (575 Lab, Dataiku) presented Kiji Inspector, an open-source tool that applies sparse autoencoders to decision tokens to surface why LLM-based agents choose particular tools. The demo covered causal ablation, layer activation extraction using patched vLLM engines, and production monitoring workflows for engineering teams.

5.0

COMPANIES · 1 SOURCE · InfoQ AI, ML & Data Engineering

Walmart Global Tech deploys multi-agent LLM batch system for inventory forecasting

At QCon AI, Aditya Mulik of Walmart Global Tech described deploying multi-agent LLM workflows for nightly batch inventory forecasting across roughly 4,700 U.S. stores, using techniques like MCP grounding and loop agents to handle schema violations, cross-department data mismatches, and network drops. He reported shifting from PTUs to pay-as-you-go inference reduced cloud inference costs from over $100k to around $10k per month.

6.0

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Advait Patel presents a playbook for zero-trust AI agent systems at QCon AI

Advait Patel (Senior SRE at Broadcom) presented a practitioner playbook at QCon AI on securing production AI agent systems: moving defenses from prompts to the tool layer, enforcing scoped workload identities, mitigating prompt-injection via tool output controls, and building audit-ready observability and kill switches to protect sensitive data while maintaining fast deployment. The talk draws on real incident lessons and existing cloud/compliance practices (AWS/GCP, SOC 2/ISO).

6.0

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Thoughtworks' Nimisha Asthagiri presents 'Beyond the Harness: Canvas for AI Agents'

At QCon AI, Nimisha Asthagiri (Thoughtworks) outlined how applying product engineering and architecture principles can create a bounded canvas for AI agents, giving them bounded autonomy, continuous drift detection, and decision records to help scale resilient software development. The talk frames this approach as an evolution from harness engineering and highlights risks like cognitive debt and the need for persistent decision traces.

4.0

RESEARCH · 1 SOURCE · InfoQ AI, ML & Data Engineering

Momento CEO Khawaja Shams presents 'Serving LLMs at Scale: The Hidden KV Cache Advantage' at QCon AI

Khawaja Shams (Co‑Founder & CEO of Momento) gave a QCon AI presentation describing distributed-systems approaches to scale LLM inference efficiently, focusing on GPU utilization, KV cache memory cliffs, tiered storage, tail-latency reduction, and disaggregated prefill/decode architectures. He framed these techniques as ways to improve fleet performance, cost economics, and availability for production LLM deployments.

5.0

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Actors for Agents — actor-model patterns for production AI (QCon AI)

At QCon AI, Manju Rajashekhar (founder & CEO of Mad Labs) presented “Actors for Agents,” arguing that the actor model maps naturally to the runtime needs of production AI agents. He described patterns — state ownership, message passing, write-through state, and supervision boundaries — to manage failures, continuations, and mutable execution graphs so agentic demos can become resilient production systems.

4.0

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Apple engineers present scaling framework-agnostic AI agent infrastructure with Ray at QCon AI

At QCon AI, Apple engineers Deepak Mohanakumar Chandramouli and Bhumik Vinodkumar Thakkar described moving AI agents from local FastAPI prototypes to resilient production deployments on Kubernetes using Ray, covering RayCluster topology, stateful memory handling, autoscaling for dynamic traffic, control‑plane designs, multi‑region setups, safety guardrails, and an LLM‑as‑a‑judge evaluation pipeline. They illustrated these operational patterns with a trip‑planner agent example and shared production learnings and observability considerations.

4.0

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Kilo Code presents large-token dataset on how developers adopt AI agents (25T cited)

At QCon AI, Brendan O'Leary of Kilo Code used real-world gateway usage data — the talk's title cites 25 trillion tokens while he also references ~40 trillion tokens — to argue developer adoption of agentic tools depends more on building trust than on model benchmark scores. He outlined a four-rung "trust ladder" and identified technical failure points in context construction, multi-model routing, and feedback loops that hinder scaling agentic workflows.

5.0

MONEY · 1 SOURCE · InfoQ AI, ML & Data Engineering

CloudZero CTO Erik Peterson at QCon AI: “You Can't Trust Your AI Bill”

At QCon AI, CloudZero founder and CTO Erik Peterson warned that AI spending is beginning to eclipse traditional cloud bills, described costly behaviors like 'tokenmaxxing' (which he estimates at roughly $5–15B), and outlined how unit economics and incentive structures are creating an engineering cost crisis; he also predicted fully autonomous AI will write, deploy, and govern code by 2028.

5.0

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Spotify presents Ads AI multi-agent production patterns using Google ADK Java

At QCon AI, Pratik Rasam (Senior Engineer, Spotify Ads) described how Spotify Ads Manager’s Ads AI runs production-grade multi-agent systems using Google ADK Java and Vertex AI on GCP. The talk covered architectural patterns, domain ownership (one agent = one package/owner), deterministic guardrails, tracing-based evaluation, and lessons on agent boundaries, tool schemas, and cost management; Spotify reported that since launch >70% of ads use AI, with ~20,000 creatives for over 7,000 advertisers.

5.0

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Elastic presents a reusable, production-grade evaluation framework for agentic AI

At QCon AI, Susan Chang (Principal Data Scientist at Elastic) described how Elastic moved from siloed, ad-hoc agent evaluations to a unified production-grade evaluation framework that combines LLM-as-judge and deterministic rules, connects Python data-science evals with TypeScript production code, and uses deep tracing to detect regressions across RAG and cybersecurity workloads while preserving domain context.

5.0

COMPANIES · 1 SOURCE · InfoQ AI, ML & Data Engineering

DoorDash details architecture and lessons from building an internal GenAI platform

At QCon AI, DoorDash engineers Swaroop Chitlur and Siddharth Kodwani presented the company’s journey building an internal GenAI platform for ~5,000 internal users, covering architectural bets, the shift from vendor-first to open-weights models, LLM and agent gateway design, and trade-offs between accuracy, latency, and cost. They emphasized API-/SDK-first design, focusing on product engineers as customers, embedding best practices, and prioritizing business impact over prototypical chatbots.

5.0

COMPANIES · 1 SOURCE · InfoQ AI, ML & Data Engineering

Forter presentation: how it trained 200 R&D staff to build AI agents in two weeks

At QCon AI, Forter principal engineer Ben Maraney described a two‑week internal sprint that gave ~200 R&D staff hands‑on experience building AI agents. The team lowered friction by running a custom MCP server, combining no‑code and code platforms, avoiding complex RAG setups, and coordinating with security and legal to speed adoption.

5.0

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Mallika Rao (QCon AI): Adaptive Recommenders in the Real World — Inference, Evals, and System Design

Mallika Rao, former engineering leader at Twitter, Walmart, and Netflix, presents at QCon AI about the operational and system-design challenges of adaptive recommendation systems. She emphasizes real-time feedback loops, retrieval freshness, multi-stage orchestration, and end-to-end latency and cost trade-offs required to continuously learn and evolve recommenders in production.

4.0

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

OpenAI's Vinoth Govindarajan presents 'Agent Harness' principles for production AI agents

At QCon AI, OpenAI engineer Vinoth Govindarajan presented 'The Agent Harness', arguing that many production agent failures stem from engineering and state-management issues rather than model hallucinations. Using case studies such as OpenClaw, he outlined core principles for reliable agent harnesses: explicit state ownership, serialized concurrent state mutations, scoped execution authority, and validating actions at the user-visible edge.

6.0

COMPANIES · 1 SOURCE · InfoQ AI, ML & Data Engineering

LinkedIn builds Contextual Agent Playbooks — organizational context layer for AI agents using MCP

At QCon AI, LinkedIn engineer Ajay Prakash presented 'Contextual Agent Playbooks and Tools', an organizational context layer built on the Model Context Protocol (MCP) that supplies procedural memory, code search and runbooks directly to coding agents. LinkedIn says the system (used across hundreds of workflows) yields a 20% productivity boost with no loss in reliability and discussed its architecture and operational guardrails.

7.0