Andrei Gudkov

Software architect/researcher (C++/Java/Python, Hadoop/Spark, ML)

LinkedIn: linkedin.com/in/andrei-gudkov/

I create high performance server and data analysis software. My 10+ year experience includes full-text search engines (Go.Mail.Ru, Nigma), hierarchical database (Deutsche Bank), a number of Silicon Valley startups. I rely on the following tech stack: C++/Linux for performance-critical components, Java/Hadoop/Spark for heavy dataflows, and Python for machine learning.

Some examples of the projects which would be relevant to me: search engines, database engines, text/image/video analysis, CDNs, ad platforms, trading robots. I prefer positions based on long-term commitment, ownership and research. High-tech startups are welcome as well.

I live in Moscow, Russia. Both on-site and remote work (UTC+3) are possible. Full time only.

Important! Please, be sure to check my technical blog. It demonstrates desired experience level and values by example.

Skills: big data, distributed computing, storage, data mining, machine learning, 24x7 online services, full-text search, web crawling, multithreading, vectorization, networking; project ownership.

Toolset:

C++, Boost, protobuf, rapidjson, OpenCV, libavcodec, flex, bison, x64/ARM SIMD, arrow, parquet, roaring, openssl, cmake, gtest
Java, Hadoop, HBase, Spark, Kafka, nio, JNI, jdbc, fastutil, jetty
Python, NumPy, scikit, xgboost, TensorFlow, matplotlib
Linux A-Z

Non-production experience: Altera Cyclone FPGA/SoC.

Experience

I work for bigdata and AI startups where I am responsible for core components. Some interesting projects:

Lighthouse AI — smart security cameras with robust object recognition

Separate GPU nodes, data pipeline latency optimization, audio subsystem, metrics collection/monitoring.

DCEngines — low-level, distributed data store integrated with Apache Spark

String/substring indexes, parquet-based storage format, cluster API, unit and integration testing frameworks.

I also write job-related articles in my spare time (see my blog).

Worked on hierarchical distributed database used for risk management.

Created embedded database client
Implemented a number of core features in a database, e.g. reference counting GC
Latency/throughput/memory optimization of everything at all levels

Mail.Ru Group is the leading Internet company in Russia. It operates public search engine go.mail.ru among other services, which served 30 millions search queries daily as of 2013.

Was responsible for development of all vertical search services. Worked in close collaboration with PM, colleagues from UI, QA, devops and other specialized groups. Major group achievements:
- Created news vertical (1000+ sources, clustering, historical data)
- Created people search over 300 million social network profiles
- Created brand new version of video search
- Added 15+ types of non-organic results and rich snippets to web search: goods, images, maps, music, news, video, weather, etc
- Created SERP construction layer. Acts as a single point of building final SERP from large number of downstream sources.
Personally made further improvements to image search. User session analysis. Centralized dataflow in hadoop.
Owned 100 online servers with highly heterogeneous roles
Led team of 7 top-quality developers, conducted technical interviews
Product screenshots

Owned image search (2,000,000 queries daily).

Restored, significantly reworked and extended image search.
Achieved big improvements in size (300M effective db size) and quality of the dataset, ranking, latency, reliability, architecture (sharding, thumbnail caching).

Nigma.Ru was a Russian metasearch engine with a variety of additional search and information services. At its peak, it had 20 employees and processed 1% of search queries in Russia.

Created music search engine over audio files accessible on the web. Complete dataflow and backend: robot, storage, indexer, ranker, search engine. Crawl set of 1 bln urls.
Created bittorrent search.
Created set of components reused by adjacent projects: robots, parsers, full-text search engine
Led team of 3 developers, conducted technical interviews (devs and sysadmins)
Product screenshots

Education

Faculty of Computational Mathematics and Cybernetics
University degree (5 years)