FissionLabs

10/12/2026

We’re proud to share that Fission Labs has earned the AWS Agentic AI Services Competency.

Alongside our AWS Generative AI Services and Generative AI Solution competencies, this further bolsters our proven track record of taking GenAI and agentic AI from pilot to production.

Whether you’re still looking for the right use case, ready to run a pilot, or scaling to production, we’re here to help you discover, pilot, and scale AI.

Thank you to our clients, our engineers, and our AWS partner teams for making this possible.

10/08/2026

Model capability is no longer what holds enterprise AI back. Every team can access a frontier model.

What separates a working system from a demo is everything around the model:
• Choosing use cases with a measurable P&L line
• Redesigning the workflow, not just adding a chat window
• Connecting AI to enterprise data and systems of record
• Security, governance, and audit trails
• Adoption, change management, and scaling beyond the pilot

That is where the Fission Labs GenAI practice works: frontier models, combined with our engineering and your domain knowledge.

See how we prioritize AI use cases: https://www.fissionlabs.com/agentic-ai-use-case-prioritization-framework?utm_source=linkedin&utm_medium=organic_social&utm_campaign=genai_practice_oct26&utm_content=p02

10/07/2026

𝗛𝗼𝘄 𝗺𝘂𝗰𝗵 𝗮𝗰𝗰𝘂𝗿𝗮𝗰𝘆 𝗱𝗼 𝘆𝗼𝘂 𝗿𝗲𝗮𝗹𝗹𝘆 𝗹𝗼𝘀𝗲 𝘄𝗵𝗲𝗻 𝘆𝗼𝘂 𝗾𝘂𝗮𝗻𝘁𝗶𝘇𝗲 𝗮𝗻 𝗟𝗟𝗠?

We benchmarked 𝗟𝗹𝗮𝗺𝗮 𝟯.𝟭 𝟴𝗕 𝗜𝗻𝘀𝘁𝗿𝘂𝗰𝘁 across FP16, FP8, FP4, and NVFP4 using vLLM and llm-compressor, evaluating performance on MMLU-Pro and MATH-500.

Our findings show that quantization is not a one-size-fits-all trade-off. 𝗖𝗮𝗹𝗶𝗯𝗿𝗮𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝘄𝗼𝗿𝗸𝗹𝗼𝗮𝗱-𝘀𝗽𝗲𝗰𝗶𝗳𝗶𝗰 𝗱𝗮𝘁𝗮 𝗰𝗮𝗻 𝘀𝗶𝗴𝗻𝗶𝗳𝗶𝗰𝗮𝗻𝘁𝗹𝘆 𝗶𝗻𝗳𝗹𝘂𝗲𝗻𝗰𝗲 𝗮𝗰𝗰𝘂𝗿𝗮𝗰𝘆.

Read the full analysis to see what the results mean for production LLM deployment.

👉 Read the blog: https://www.fissionlabs.com/blog-posts/llm-quantization-accuracy

Large language models are expensive to run. An 8 billion parameter model, in case of Llama 3.1 8B Instruct needs around 16 GB of memory just to hold its weights in BF16, before you even account for activations, KV cache, and batching overhead. Scale that to multiple GPUs serving thousands of request...

10/06/2026

𝗔𝗜 𝗮𝗱𝗼𝗽𝘁𝗶𝗼𝗻 𝗶𝘀 𝗮𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗶𝗻𝗴. 𝗕𝘂𝘁 𝘁𝘂𝗿𝗻𝗶𝗻𝗴 𝗮𝗱𝗼𝗽𝘁𝗶𝗼𝗻 𝗶𝗻𝘁𝗼 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝘃𝗮𝗹𝘂𝗲 𝗿𝗲𝗺𝗮𝗶𝗻𝘀 𝗮 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲.

McKinsey’s 2026 research found that 44% of organizations report scaling AI across the enterprise, up from 38% a year earlier. Yet the share reporting a positive EBIT impact held flat at 37%

The gap between adoption and scale is rarely about access to better models. It is about everything around the model: 𝗱𝗮𝘁𝗮 𝗿𝗲𝗮𝗱𝗶𝗻𝗲𝘀𝘀, 𝗶𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻 𝘄𝗶𝘁𝗵 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝗼𝗳 𝗿𝗲𝗰𝗼𝗿𝗱, 𝗴𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲, 𝗼𝗯𝘀𝗲𝗿𝘃𝗮𝗯𝗶𝗹𝗶𝘁𝘆, 𝗮𝗻𝗱 𝗰𝗼𝘀𝘁 𝗰𝗼𝗻𝘁𝗿𝗼𝗹.

Over the next four weeks, Fission Labs will explore what it takes to move enterprise 𝗔𝗜 𝗳𝗿𝗼𝗺 𝗽𝗶𝗹𝗼𝘁 𝘁𝗼 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻, with the engineering foundations needed to scale with confidence.

Start by identifying where your AI use cases stand:
https://www.fissionlabs.com/agentic-ai-use-case-prioritization-framework?utm_source=linkedin&utm_medium=organic_social&utm_campaign=genai_practice_oct26&utm_content=p01

09/28/2026

We’re excited to announce that Fission Labs is now part of the Claude Partner Network.

This partnership brings together Anthropic’s AI capabilities with our expertise in AI, GenAI, Agentic AI, and enterprise engineering to help organizations move from identifying the right opportunities to building and deploying production-ready solutions.

From use case discovery and prioritization to pilots and production, we help enterprises turn AI potential into measurable business impact.

09/24/2026

What happens when AI agents need access to information beyond their training data?

We explored web search with Amazon Bedrock AgentCore to see how agents can retrieve real-time information and use it to generate more relevant responses.

Read our technical walkthrough to see the implementation and key learnings.

🔗 https://www.fissionlabs.com/blog-posts/amazon-bedrock-agentcore-web-search-test

We ran six tests on Amazon Bedrock AgentCore Web Search: determinism, the 200-character limit, phrasing sensitivity, and what it means for agent design.

09/11/2026

We’re hiring at Fission Labs!

We’re growing our team and looking for talented professionals across QA, Salesforce, Finance & Commercial Operations, and Inside Sales.

Open roles include:
• Automation QA (Python + Playwright) | 3–5 years | Hyderabad
• Salesforce Marketing Cloud Developer | 5+ years | Remote
• Finance & Commercial Operations Lead | 7–11 years | Hyderabad
• Inside Sales Manager | 7–11 years | Hyderabad
• Inside Sales Executive | 3–5 years | Hyderabad

If you’re looking for your next opportunity and your experience matches any of these roles, we’d love to hear from you.

📩 Send your resume to [email protected]

Know someone who could be a great fit? Share this post with them.

09/08/2026

𝟵 𝗼𝘂𝘁 𝗼𝗳 𝟭𝟬 𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝗺𝗲𝗻𝘁𝘀 𝗽𝗮𝘀𝘀𝗲𝗱. 𝗧𝗵𝗲 𝗼𝗻𝗲 𝘁𝗵𝗮𝘁 𝗱𝗶𝗱𝗻'𝘁 𝗲𝘅𝗽𝗼𝘀𝗲𝗱 𝗮 𝗿𝗲𝗮𝗹 𝗯𝗹𝗶𝗻𝗱 𝘀𝗽𝗼𝘁.

We ran AWS Security Agent's Design Review on a serverless AWS architecture using CloudFront, S3, API Gateway, Lambda, Cognito, and DynamoDB.

The result? 9/10 compliant.

The exception was Log Protection: “insufficient data.”

Our design covered encryption, access control, retention, and tamper prevention. But it didn't answer one critical question:

Was PII being redacted before it reached the logs?

Encrypting your logs doesn't help if a Lambda function writes a raw request containing sensitive data for debugging.

A good reminder that security controls can look solid individually while still leaving gaps between them.

We documented the finding, remediation guidance, and what we learned here:

https://www.fissionlabs.com/blog-posts/aws-security-agent-design-review-real-findings-inside

We tested AWS Security Agent Design Review on a serverless AWS app. Here is the PII logging gap it caught and what the feature costs today.

09/01/2026

We’re proud to be recognized by AWS for our DevOps expertise, helping businesses move faster, improve reliability, and scale with confidence.

Thank you to our AWS Partner team and the Fission Labs team for their continued support and collaboration.

08/28/2026

We're excited to announce that Fission Labs is now an AWS 𝗘𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲-𝗕𝗮𝘀𝗲𝗱 𝗔𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗶𝗼𝗻 (𝗘𝗕𝗔) 𝗖𝗲𝗿𝘁𝗶𝗳𝗶𝗲𝗱 𝗣𝗮𝗿𝘁𝗻𝗲𝗿!

This certification reflects our deep expertise in AWS and our commitment to driving real business outcomes. Through hands-on, outcome-focused engagements, we help organizations accelerate their agentic, AI transformation, cloud, data modernization journeys.

As an EBA Partner, we're empowered to collaborate even more closely with AWS to:

✅ AI and innovation
✅ Align on strategic outcomes
✅ Build organizational and technical capability
✅ Accelerate delivery of measurable results

A big thank you to our amazing team, clients, and AWS partners who continue to inspire and support our journey.

Here's to accelerating innovation and building what's next together!