OmniTier

Milpitas, CA & Rochester, MN

  • Coding principles
    Hacky prototype before clean code, Bleeding-edge stack
  • Modern Tech Stack
    React, Go, Python, JavaScript, Java, C++, C

Our Mission

Memory-Centric, High Performance Computing Solutions for the Life Sciences. Our vision is to turn servers into supercomputers for prescribed applications, at a fraction of the cost, and deliver those products with the best customer support. Whole genome sequencing informatics pipelines achieve greater accuracy and faster run-times.  Affordable terascale analytics using the same tiered memory architecture.  We develop solutions for bioinformatics, scientific computing, and web services. Our customized solutions take full advantage of new flash memory resources (tiers) which expand computational storage. Run times are dramatically reduced while new insight is gained from big data. - Handle bigger datasets achieving higher accuracy in your genome sequencing, AI training, and analytics applications. - Speed up time-to-actionable results in genome sequencing; AI pipelines; response times in web services and NoSQL data base query applications. - Reduce infrastructure costs for your applications. OmniTier’s technology and solutions enable all three needs for big and fast data applications: higher accuracy, reduced latency, and lower costs.

Tech Stack

CompStor is a product family of accelerated whole genome sequencing (WGS)  informatics pipelines and Analytics Tools (AI, machine learning, matrixes, and graphs). It is provided as a server cluster in either 2 or 4 node configurations with optimized data ingress and egress. CompStor brings Computation to Storage. Our initial CompStor NovosTM product is a WGS variant caller featuring both de novo assembly and reference alignment based methodologies. It produces highly accurate variant calling with accelerated run-times. MemStac is a caching solution that drives performance in web services and edge data analysis. DRAM based solutions are expensive and lack storage capacity for performance gains. MemStac delivers an optimized DRAM/NVMe-SSD solution with performance gains and cost savings from day 1. Designed for a tiered memory architecture utilizing NVMe SSDs and small amounts of DRAM, MemStac is capable of exceeding 7 million cache requests per second and supporting up to 4 terabytes cache capacity per server node at a fraction of the cost of today’s DRAM-only solutions. MemStac is fully Memcached compliant and can be integrated into existing cache clusters seamlessly, requiring no changes to the existing application (client) libraries. Using select off-the-shelf NVMe SSDs as cache memory within a standard Intel dual-socket server, MemStac delivers native SSD performance at 3.7M operations per second (100 percent Get operations) and 3.2M operations per second (80 percent Get operations / 20 percent Set operations) for 100 byte records. With widely-used workloads, this performance increases to an industry-leading 6.1M and 4.7M operations per second, respectively, using DRAM storage at one-eighth of the cache capacity. These levels of performance are achieved with less than 1 millisecond average latency. In addition, MemStac seamlessly delivers 10GbE network-limited throughput with typical workloads exceeding 250 bytes in average size, similar to the DRAM-only open-source solutions. To achieve these technological breakthroughs, our team has deep domain expertise in: - multi threading - concurrency - parallel processing - fault tolerance - data replication - large scale distributed compute and storage systems

What Our Team is Reading

considering building a small app in clojure to try it out https://www.demystifyfp.com/clojure/blog/clojure-in-production/


Team Q&A

  • Tiered-Memory Memcached Solution for Database Caching

    OmniTier

  • How do you integrate MemStac with Spring and Hibernate?

    OmniTier


Vacation & Benefits

6-week PTO.

Culture

  • Always be learning
  • Psychological safety

Cool Tech

  • Whatever c++ people call functors
  • Open source contributor

Day in the Life

  • Private offices
  • Great for parents

Team members

  • 20 people

Fullstack Software Engineer

Milpitas, CA・New Grad・Full Time

Apply

More Companies You May Like

3 jobs

Envato

Our story starts in 2006 with three founders in a Sydney garage (no, we’re not kidding). Born from a desire to earn a living doing what they loved, with the flexibility to do it from anywhere, Enva...

Melbourne, Australia

Faraday

Faraday is the end-to-end machine learning platform for the B2C stack. Consumer data, algorithms, visualizations — it’s all built in and ready to power your teams and systems in 6-8 weeks....

Burlington, VT
2 jobs

BuildZoom

Transforming the $1.3 trillion dollar old-school construction industry using data, natural language processing, and machine learning. We are improving the remodeling experience for hundreds of mill...

San Francisco, CA