Yesterday, I attended the Code on JVM meetup, where industry experts and experienced professionals delivered three insightful technical sessions. The event covered the following topics:
- DSA in Production – A Case Study
- Uncovering Ideas Worth Building in the AI Era
- Streaming Large File Downloads in Java
Each session provided practical insights, real-world experiences, and valuable knowledge on applying these concepts in the software industry. In this blog, I'll share the key takeaways and lessons I learned from each session.
1. DSA in Production – A Case Study
Taken by - Hareesh Rajendran (Founder & CEO, Hayan Tech)
This session demonstrated how data structures and algorithms are applied in large-scale production systems rather than just coding interviews.
Key Takeaways
Bloom Filters
- A probabilistic data structure used to quickly check whether an element may exist.
- Extremely memory-efficient.
- Used in databases and caching systems to avoid unnecessary lookups.
HyperLogLog
- Used for estimating the number of unique elements in massive datasets.
- Requires very little memory.
- Based on probability and mathematical concepts.
Common use cases include:
- Counting unique Reddit users
- Google Analytics
- Analytics dashboards
LSM Trees (Log-Structured Merge Trees)
Optimized for high write performance.
Widely used in databases handling large volumes of data.
Consistent Hashing
Helps distribute data across multiple servers.
Reduces data movement when servers are added or removed.
Commonly used in distributed systems and load balancing.
HNSW Graph (Hierarchical Navigable Small World)
HNSW is a graph-based data structure that enables extremely fast similarity searches among millions of vectors. It is widely used in AI applications such as semantic search, recommendation systems, vector databases, and Retrieval-Augmented Generation (RAG).
Skip List
A Skip List is a probabilistic data structure that provides fast search, insertion, and deletion operations by using multiple linked levels. It offers performance similar to balanced trees while being simpler to implement and is commonly used in Redis and database storage engines.
Hexagonal Grid (H3)
H3 is a geospatial indexing system that divides the Earth's surface into hexagonal cells for efficient location-based processing. It is widely used by mapping, ride-sharing, and food delivery platforms to optimize routing, nearby searches, and geographic analysis.
Takeways:
Totally about the data structure how it works, why it works by knowing this easyily understand the concept and above are DSA algorithm used in many use cases and case study.
2. Uncovering Ideas Worth Building in the AI Era
Taken by - Agnel John D (Founder & CEO of Error makes clever)
This session focused on validating ideas before investing time in building products.
Lessons Learned
- Building a product is easier than finding people who truly need it.
- Validate your idea by talking to potential users first.
- Identify customers who are willing to pay.
- Your first paying customers are the strongest proof that your idea solves a real problem.
- Once validation is complete, commit fully and execute with focus.
Takeways
How to use AI efficiently
By using AI to build model
How it works for working professions
Then how prioritize the product to the market using AI
Session 3: Streaming Large File Downloads in Java
Taken by - Kalaivani Mohan (Software Engineer, IBM)
The final session explained how streaming large file downloads improves application performance. Instead of loading an entire file into memory, Java streams the file in smaller chunks, reducing memory usage and enabling faster, more scalable downloads. This technique is widely used for serving videos, PDFs, reports, cloud storage files, and other enterprise applications, making backend systems more efficient and reliable.
Takeways
This session is fully about the large file downloads for users. By split as chunk to n number of users with the help of Java.
Final Thoughts
This meetup was an excellent learning experience. It highlighted how computer science concepts like Bloom Filters, HyperLogLog, LSM Trees, and Consistent Hashing power modern applications, while also emphasizing that successful AI products start with solving real customer problems—not just writing code.
A big thanks to the speakers and organizers for sharing their practical experiences. Looking forward to attending more such sessions!
#CodeOnJVM #Java #DSA #SystemDesign #AI #Software engineer
Top comments (0)