Skip to main content

2 posts tagged with "conferences"

View All Tags

10 takeaways from xconf 2024

· 3 min read
D Balaji
D Balaji
Lead Design Technologist

XConf, the premier tech conference hosted by Thoughtworks, was held on November 22, 2024, at Marriott Whitefield, Bengaluru. The event was accessible through registration using a professional email at Thoughtworks XConf.

Event Highlights​

Despite starting slightly behind schedule and facing time management challenges for speaker sessions, the event was vibrant and informative. Sponsored by AWS and CockroachDB, it featured five distinct booths:

  1. AWS
  2. CockroachDB
  3. Thoughtworks Careers
  4. Thoughtworks Immersive Experiences in the Metaverse
  5. Thoughtworks Publications

The event was structured into three segments:

  1. Common Sessions: Keynotes and talks by AWS and other industry leaders.
  2. Specialized Tracks: Focused sessions across specific themes.
  3. Workshops: Exclusive, registration-based, hands-on workshops.

The keynote speakers were particularly engaging, setting an inspiring tone for the day.

Themes for Specialized Tracks​

  1. Machine Learning, Data, and AI
  2. Distributed Systems
  3. Emerging Technologies: Including SDV, XR, and Embedded Systems

A standout moment was a fascinating talk by the Director of the Indian Astrophysics Department, highlighting the role of technology in space exploration. Interestingly, Thoughtworks has collaborated with the department for their software needs.

Key Takeaways​

Here are 10 notable insights from the conference:

  1. Software Development as a Team Sport

    • AI assistants should enhance the entire software development lifecycle rather than support isolated coding efforts.

  2. From 10x Developers to 10x Teams

    • The aim of AI is to empower teams, fostering collaborative processes and tools for impactful delivery.

  3. AI Across the Software Development Lifecycle

    • Beyond chatbots, AI is revolutionizing software processes, including research, planning, design, testing, deployment, and maintenance.

  4. AI Artifacts for Enhanced Productivity

    • Sharing generative AI (GenAI) prompts across teams can significantly boost efficiency, supported by tools like Haiven Team Assistant.

  5. Observability 2.0

    • Innovations like canonical log lines are scaling observability practices while reducing network loads.

  6. GenAI for Legacy Code Understanding

    • GenAI facilitates reverse engineering legacy code, enabling seamless tech migrations.

  7. Rethinking Codebase Documentation with GenAI

    • GenAI excels in generating documentation, capturing module links, and documenting architecture, epics, and stories.

  8. AI Tools for Diverse Problem-Solving

    • Utilize GenAI prompts for code understanding, RAG (Retrieval-Augmented Generation) for problem-solving, and Graph + RAG for capability analysis in codebases.

  9. Local-First Software Development

    • A paradigm emphasizing on-device computation for enhanced privacy, security, and real-time AI inferencing.

  10. Evaluating LLM Performance

    • Techniques like "eval" and "vibe checking" are emerging for benchmarking LLMs, with both self-assessment and human validation improving model efficiency.

Additional Perks​

  • Meet and interact with authors of Thoughtworks publications.
  • Opportunities for networking, paired with great coffee and exclusive goodies.
  • A focused event highlighting use cases in enterprise software.

In conclusion, XConf 2024 provided a dynamic platform for exploring cutting-edge tech trends, fostering meaningful collaborations, and envisioning the future of enterprise software development.

Google GenAI Developer Day Bangalore 2024 Learnings

· 3 min read
D Balaji
D Balaji
Lead Design Technologist

Gen AI Dev Day

This blog post summarizes key takeaways from the Google GenAI Developer Day in Bangalore, focusing on technical aspects relevant to developers.

1. GenAI Impact and Demand​

While "AI" is often misused in marketing, the conference solidified the transformative power of Generative AI (GenAI) across personal, professional, and societal applications. For businesses, GenAI tools enhance developer productivity by:

  • Auto-completion: Generating code snippets based on prompts.
  • Context-aware Test Generation: Automatically creating test cases.
  • Code Explanation and Review: Providing explanations and assisting with code reviews.
  • Multilingual Support: Supporting various programming languages and transpiling between them.
  • Full Codebase Awareness: Enriching analysis, refactoring, and optimization.
  • Efficient Code Generation: Capable of generating substantial code volumes (e.g., 100k lines) from single prompts.

2. Code Enhancement Use Cases​

Live demonstrations showcased GenAI's capabilities. Notably, the creation of a shopping cart app with star ratings on both frontend and backend required only a few prompts.

VS Gemini vs. Code Assist Addon

While both tools generate code, Gemini offers a key advantage: it can learn and generate code for new or less common libraries on-the-fly, surpassing limitations of pre-trained languages.

Google Code Assist seamlessly integrates with Github, GitLab, and Bitbucket, generating code for diverse use cases. For enterprise security, models are trained on single-tenant private clouds, garnering trust from financial institutions.

Developer Tools and Resources

  • Vertex AI and Natural Language Integration: Enables log and crash analysis using natural language queries within GCP.
  • Code Assist Availability: Extends to various IDEs like Android Studio, GCP, and databases.
  • AI Practice Platforms: Google Colab for experimentation and Google Codelabs offering industry-leading examples.
  • GenAI Database Support: AlloyDB delivers high-performance, fully-managed database capabilities specifically suited for GenAI. Similar capabilities were showcased by MongoDB, Neo4j, and Elasticsearch.
  • Vertex AI Multimodal Exploration: Explore GenAI functionalities for various modalities within GCP via "Goto Vertex AI > Multimodal > Ask Prompt".
  • Google Cloud Functions Update: Google Cloud Functions are now rebranded as Google Cloud Run.

4. Foundational Concepts for GenAI Development​

Developing GenAI applications requires:

  • Training Data: To build robust models.
  • Context-Aware Responses: To generate relevant outputs based on natural language prompts.
  • Unified Data Handling: Images, videos, and text are treated as equally compatible data sources.

5. Retrieval-Augmented Generation (RAG)​

While details weren't extensively covered, the session briefly introduced RAG as a technique for enhancing GenAI models.

6. Hackathon​

The concluding session featured a hackathon based on the Codelabs.