Our contribution to the field and the community

We help new generations of mathematicians and programmers unlock their potential, as others once helped us: we support universities, make open courses, run competitions and lectures

Open courses on Stepik

40+ hours of video lectures and 180+ lessons in university-level courses: algorithms, parallel programming, reinforcement learning, Unix. More than 10,000 students have taken the courses. Partner universities use them in their curricula

  • Algorithms and data structures

    5.0 / 56300+ learners

    From models of computation to search trees: sorting, dynamic programming, greedy algorithms, graphs, segment tree

    Teacher: Maria Gordenko, HSE University

    67 lessons, 9 h of video, 149 problems

  • Parallel programming

    4.9 / 51800+ learners

    Threads and IPC, synchronisation primitives and algorithms, memory models, lock-free stack and queue, RCU, profiling

    Teacher: Evgeny Kalishenko, St Petersburg University, HSE University, ITMO University

    52 lessons, 13 h of video, 14 tests

  • Deep reinforcement learning

    4.4 / 51400+ learners

    Reinforcement learning basics, classic Deep RL algorithms and an overview of subfields, practice on classic problems. Python, ML

    13 lessons, 11 h of video

  • Introduction to Unix

    4 / 51100+ learners

    The command line, POSIX shell and scripts, regular expressions and text utilities, git, package managers and build systems

    44 lessons, 7 h of video

  • AKOS

    Coming soon, the course is in development

    Computer architecture and operating systems: assembly, memory and caches, processes and threads, system calls, networking and file systems

    Teachers: David Badalyan, Alexander Myltsev

  • Basic C++

    Coming soon, the course is in development

    C++ from zero to confident code: syntax and types, pointers and references, memory and object lifetime, classes, STL containers, basics of templates and builds

    Teacher: Olga Khvostikova

See all courses

4th cohort, September-December 2026

Effective C++ course

A course on high-performance C++ in an HFT context. We built it to train our own team, then opened it to students. It is now an elective in the programmes of HSE University, St Petersburg University and ITMO University; for students of other universities there are open places through a competitive selection

Join the waitlist

Programme: the C++ object model and memory layout, std containers, building programs and libraries, CMake, templates and metaprogramming, gdb, optimisation, CPU and memory, Linux internals, multithreading.

Over three runs the course grew from an experiment at St Petersburg universities into open enrolment. In 2024 selection went through Yandex.Contest, and 38 people passed. In 2025 we opened 100 places for everyone: 182 applications, 92 enrolled. Applicants were students of MIPT, ITMO, HSE University, Moscow State University, St Petersburg University, Novosibirsk State University, Innopolis and universities in Belarus, Germany and Korea, and with them senior engineers and entrepreneurs

FOR WHOM
2nd–4th year students who want to learn more about high-performance C++ and the HFT industry. Useful to both quants and developers if they already know the language
FORMAT
Video lectures, auto-graded problems, two manually reviewed projects, weekly Q&A, guest lectures by developers. Online, 10-12 weeks, 6-8 hours a week
AUTHOR
Alexander Tiunov, Senior Software Engineer on our team, previously taught C++ at the Yandex School of Data Analysis. He gives the lectures himself and answers questions at a Q&A every week.
SELECTION AND RATING
3+ applicants per place, average course rating 4.91 out of 5. Most participants rated their skill growth after the course higher than expected, some much higher
  • ITMO University
  • HSE University, Moscow
  • HSE University, St Petersburg
  • St Petersburg University

Contests and competitions on real problems

We like to compete and especially like to win. We carry this outside through our own and partner competitions; some of their problems come from our HFT practice

Subscribe to updates
  • Day X

    October 2, 2026

  • Low-latency Data Transfer Challenge

    Prize fund $12,000 A problem from our C++ engineers' practice: deliver a data stream from one source to several receivers with minimal latency.

    Finished, 03.07-30.08.2026

Events, meetups, technical talks and lectures

Follow announcements in Telegram
  • Zero Cost Conf 2026

    A conference on applied C++ with a focus on performance, Moscow and Belgrade. A talk by our team

    1 August 2026, finished

  • Back to Back Conf 2026

    Talk "Tracing and profiling for free": with examples from HFT. Our engineer is on the programme committee

    August 2026, finished

  • OpenTalks.AI 2026

    Talk "Machine Learning in HFT: Winning in an Ultra-Competitive Market", Belgrade

    19 February 2026, finished

  • C++ Russia 2025

    Alexander Emelenko, Senior C++ SWE: multithreaded programming in high-load systems. A full hall, almost an hour of questions afterwards.

    2025, finished

  • Practical project: hot-cold split in realtime systems

    Students write an LLVM plugin that marks hot and cold sections of code, supervised by a Senior engineer

    Autumn 2024, finished

  • Sponsorship of the ICPC NEF final and VKOSHP

    The Northern Eurasia final and the main round of VKOSHP, a semifinal of the international competition

    December 2023, finished

  • Open lectures at the Independent University of Moscow and universities

    Two areas, ML for trading strategies and low-latency infrastructure. 8 places, 2.5 months, 20 hours a week

    Spring 2023, finished

  • Internships in HFT projects

    Two areas, ML for trading strategies and low-latency infrastructure. 8 places, 2.5 months, 20 hours a week

    Spring 2023, finished

  • Online lecture "Quantitative Research

    How mathematics and programming help win in financial markets". An open lecture for students about a quant's work in HFT

  • Open lectures at the Independent University of Moscow and universities

    What problems engineers and researchers solve in HFT and why mathematics, programming, data work and low-level optimisation are needed here

    Finished

We work inside the curriculum

Partner universities where our courses run: HSE University, St Petersburg University, ITMO University. For more than four years we have supported the Independent University of Moscow: we cover more than 90% of its financial needs. We do not come with a one-off lecture and leave: our formats are built into the curriculum and repeat every year.

  • A course in the curriculum

    Our courses run at partner universities as part of the curriculum. A student takes the course and gets a credit at their own university.

  • Practical problems for students

    We give problems from our practice. A student takes one as a term paper or a research project, supervised by an engineer who solves this problem at work.

  • Lectures and competitions

    Lectures by engineers on how the work is done inside. Optimisation contests on real problems from our practice, with a prize fund and a shortened hiring track.

  • Independent University of Moscow (MCCME)
  • HSE University, Moscow
  • HSE University, St Petersburg
  • St Petersburg University
  • ITMO University
  • ICPC

4+ years as the main sponsor of the Independent University of Moscow

Telegram channel with 4000+ subscribers

Breakdowns of production problems, talk notes, course and competition announcements — in the Telegram channel

All articlesTelegram channel

For coaches, schools, universities and competition organisers

We work with those who prepare strong school and university students. Tell us what you need, and we will discuss the format: a course in the curriculum, a joint competition or support for your project.

«I grew up in competitive programming. I did not become a programmer, but this experience helps me in almost everything I do. Now I lead spectral:future, the area that supports strong school and university students in mathematics and IT. Spectral invests in this systematically, because we think it right to support the community we grew out of ourselves »

Anatoly Fedchenko

Head of spectral::future

afedchenko@spectral.tech

t.me/avfedchenko

Contact

How to get into Spectral

  • Apply for a vacancy

    The standard track takes about three weeks; if needed, we complete it in a few days

    Vacancies
  • Through competitions

    We offer the best competition participants a shortened interview track

  • Through university projects

    Study projects, problems for students, courses by the company's engineers

Questions and answers

  • Is experience in olympiads, competitive programming or competitions required?

    Olympiad experience is not required, but it correlates positively with the skills we need. Other examples of achievements and a successful track also work: the top of the ranking at a strong technical university, scientific achievements, a fast growth track, successful projects at work.

  • Do you hire people without experience?

    For junior positions, for example Junior Quant, work experience is not required. For middle and senior, real work projects and practical skills are the key factor.

  • Can I combine working with you and studying?

    Yes. Our employees graduated from MIPT, HSE University and St Petersburg University while working. We give study leave for exam sessions and the diploma.

  • How long does the hiring process take on average, and what are the stages?

    About three weeks as a baseline, to go through all the stages on a comfortable schedule. If needed, we hold all the interviews within a few days. From time to time there are shortened tracks based on participation in our projects.

  • Are interviews held in person or remotely?

    All stages can be done remotely. Sometimes we invite you to the office for the fit interview, to meet the lead and the team.

  • What is your work format?

    In the office only: the speed of communication and the quality of teamwork are many times higher with in-person interaction. In rare cases we discuss remote work with a successful candidate.

  • How many hours a week do you work?

    Technical teams work 40-45 hours a week. Plus additional forms of leave on top of those required by law.

  • How do I recommend a friend?

    If you have friends who fit one of the open positions, write to us in Telegram and we will contact them.