A Gentle Introduction to Reinforcement Learning
A gentle introduction to Reinforcement Learning
Satwik Kansal
Mentor
(100 reviews)
US$30.00
For every 15 mins
233
Sessions/Jobs
ABOUT ME
Algorithmic Trading Systems & AI/LLM Engineer | Python
Algorithmic Trading Systems & AI/LLM Engineer | Python
I build the execution, data and backtesting infrastructure that lets a trading strategy run unattended, plus the LLM and agent systems around it, inside your repo, without needing your alpha.
Lead developer at a Swiss fintech building automated trading systems. Founder and lead developer of a finance-and-AI software consultancy. Author of a Python open-source project with ~37,000 GitHub stars, featured on the Hacker News front page four times and endorsed by Mr Andrej Karpathy, co-founder of OpenAI. Published the video course "Hands-on Reinforcement Learning with TensorFlow" with Packt.
I do not build "guaranteed profit" bots or sell signals. I build the systems that run yours safely, and I will tell you in the first fifteen minutes whether what you have is an infrastructure problem or a strategy problem.
What people book me for:
- Trading-system architecture and execution-path review: order management, partial fills and all-or-none orders, reconciliation after disconnects, persistent order state, kill switches, paper-trade-first deployment.
- Exchange and broker connectivity: WebSocket and REST integrations (Alpaca, CCXT-supported venues, Binance), rate limits, idempotent order submission.
- Backtesting infrastructure and the backtest-versus-live gap: fills, slippage and latency assumptions, replay fidelity, parameter search without overfitting.
- Market making and liquidity tooling across venues.
- LLM and agent systems for finance workflows: orchestration, evaluation pipelines, tool-use design, reinforcement-learning approaches where they actually fit.
- Python systems engineering: asyncio, PostgreSQL, FastAPI and Django, production reliability.
How a first session usually goes: you share the repository or a design, I ask three questions (budget and timeline; whether you already own the strategy, accounts and data; whether you want infrastructure or "a profitable bot"), and we spend the rest of the hour on the highest-risk part of your system. Longer work continues as a fixed-scope code-review job or an hourly engagement on this platform.
Based in Bengaluru (IST), with daily overlap with European hours and the US-East morning. I reply to trading, fintech and AI requests within minutes when online.
More at https://satwikkansal.xyz | Consultancy: https://zobyt.com
Hindi, English
New Delhi (+05:30)
Joined August 2017
EXPERTISE
8 years experience | 1 endorsement
Execution engines, order management and backtesting infrastructure for equities and digital assets. Execution and market-making systems w...
Execution engines, order management and backtesting infrastructure for equities and digital assets. Execution and market-making systems with gap-open reconciliation, crash recovery from a persistent order ledger and kill-switch controls. Exchange connectivity over WebSocket and REST (Alpaca, Binance, CCXT-supported venues). Paper-trade-first deployment with risk controls before anything touches live capital.
10 years experience | 8 endorsements
Production Python for trading infrastructure and AI systems: asyncio, FastAPI, Django, PostgreSQL, pytest, Docker. Author of wtfpython, a...
Production Python for trading infrastructure and AI systems: asyncio, FastAPI, Django, PostgreSQL, pytest, Docker. Author of wtfpython, a Python open-source project with ~37,000 GitHub stars, featured on the Hacker News front page four times. Core contributor to Python open-source projects (coala, Oppia, Italian Mars Society); Google Summer of Code student and later mentor.
8 years experience | 1 endorsement
LLM orchestration and agent workflows: built an LLM workflow engine composing multiple model types as first engineering hire at an AI-inf...
LLM orchestration and agent workflows: built an LLM workflow engine composing multiple model types as first engineering hire at an AI-infrastructure startup, before the commercial LLM boom. Evaluation pipelines and tool-use design for agents. Reinforcement learning (authored the Packt video course "Hands-on Reinforcement Learning with TensorFlow"). Applied ML for financial data, model training and evaluation, PyTorch.
10 years experience | 6 endorsements
Experienced with building backend systems in Python (Django and Flask) and nodeJS. Worked with both relational and non-relational databas...
Experienced with building backend systems in Python (Django and Flask) and nodeJS. Worked with both relational and non-relational databases in the past.
Knows about creating production-ready applications with containerized architecture, continuous integration (CI/CD), and scaling on the cloud.
8 years experience
Trading infrastructure for equities and digital assets: execution, order management, market making and backtesting systems. 3.5 years as ...
Trading infrastructure for equities and digital assets: execution, order management, market making and backtesting systems. 3.5 years as lead developer at a Swiss fintech building automated trading systems. Portfolio and risk tooling, real-time monitoring and alerts. Payments-stack engineering at one of India's largest e-commerce platforms.
8 years experience | 3 endorsements
Have experience with Data processing, EDA, model training, evaluation, and report building. Familiar with pipelined architectures used in...
Have experience with Data processing, EDA, model training, evaluation, and report building. Familiar with pipelined architectures used in Machine learning.
Authored a course called Hands-on Reinforcement Learning with Python (published on Udemy, Packt, and Oreilly's Safari).
Worked on a Space Weather Forecasting for the Italian Mars Society.
Data Science intern at Anheuser-Busch Inbev.
Contributed my learning through Machine Learning guest blogs on various platforms.
Some of my previous research topics include:
- Deep Reinforcement Learning
- Text classification (predicting sentiments, likes, etc. of tweets and Reddit posts)
- Convolutional Neural Networks (CNNs)
- Generative Adversarial Networks (GANs)
I started my Data Science journey in 2015 with Udacity's Machine Learning Nanodegree, also completed Udacity's Deep Learning Nanodegree in 2018.
8 years experience | 4 endorsements
On-chain execution systems: an MEV extraction system across Ethereum and EVM-compatible chains combining off-chain opportunity detection ...
On-chain execution systems: an MEV extraction system across Ethereum and EVM-compatible chains combining off-chain opportunity detection (DEX arbitrage, loan liquidations) with on-chain execution through smart contracts and private transaction relays, avoiding front-running and failed-transaction gas. Smart contracts in Solidity, wallet integrations, real-time streaming payments and token-distribution infrastructure, and exchange connectors for EVM-based venues.
REVIEWS FROM CLIENTS
(100 reviews)
Manu Pasricha
May 2024
Exceptional instructor. Extremely good with github.
kuan
July 2022
good
John Doe
March 2022
Satwik was very helpful in helping me understand how to tackle my issue.
Danny TJ Kan
March 2022
Satwik is a patient, polite, and strong logic person with a good communication skills!
Our team and I are looking forward to the future cooperation with him!
Thanks for the great work!
Dionte Johnson
March 2022
Great developer!
Timothy Conwell
December 2021
Satwik was a joy. He solved my problem in an efficient and timely manner. He is a very knowledgeable software engineer. I was fortunate to have found him. Highly recommended!
Ann Choe
October 2021
Satwik is very friendly, efficient, and competent. He has been my go-to mentor for a while now, and I'd recommend him to anyone who needs a good mentor to help out with their projects.
Manu Pasricha
August 2021
Satwik has a great knowledge about trading, cryptocurrency and blockchain. He is very patient and explains things in great details. Highly recommended for anything crypto, blockchain and trading related.
Manu Pasricha
August 2021
Great informative session with Satwik on blockchain and crypto. He is very knowledgeable in this field of crypto and blockchain. Highly recommend him.
Brian loughran
January 2021
Very helpful and super friendly guy, highly recommend
SOCIAL PRESENCE
wtfpython
What the f*ck Python? 😱
Python
37089
2656
python_blockchain_app
A fully functional blockchain application implemented in Python from scratch (with tutorial).
Python
851
463
1351 Reputation
4
19
33
EMPLOYMENTS
Founder & Lead Developer
Zobyt
2022-01-01-Present
Finance-and-AI software consultancy; I lead architecture and delivery.
- Trading infrastructure: a US-equities grid-trading engine with g...
Finance-and-AI software consultancy; I lead architecture and delivery.
- Trading infrastructure: a US-equities grid-trading engine with gap-open compensation, crash-resilient order management on a PostgreSQL order ledger, all-or-none fills and kill-switch controls, paper-trade first; automated market making and liquidity management across exchanges; backtesting and strategy-optimization infrastructure.
- On-chain execution: an MEV extraction system across Ethereum and EVM chains combining off-chain opportunity detection with on-chain execution through private transaction relays.
- AI systems: LLM orchestration, agent workflows and evaluation pipelines for finance use cases.
- The consultancy has delivered 200+ projects processing over $2B in volume.
Python
Trading Systems
LLM
View more
Python
Trading Systems
LLM
View more
Lead Developer - Automated Trading Systems
Hyperion Fintech AG
2020-12-01-2024-06-01
Led development of automated trading systems and trading-strategy infrastructure for a Swiss fintech, as lead developer working directly ...
Led development of automated trading systems and trading-strategy infrastructure for a Swiss fintech, as lead developer working directly with the founders: execution and monitoring, backtesting, and the operational tooling needed to run strategies unattended. Systems work throughout; strategy logic and results stayed with the client.
Python
Algorithmic trading
Trading Systems
View more
Python
Algorithmic trading
Trading Systems
View more
Senior Software Engineer - AI Infrastructure (first hire, contract)
Mantium
2021-01-01-2021-05-01
First engineering hire at an AI-infrastructure startup. Built the AI workflow engine that composed multiple model types, including large ...
First engineering hire at an AI-infrastructure startup. Built the AI workflow engine that composed multiple model types, including large language models, into pipelines, before the commercial LLM boom. Contract role.
Python
Machine Learning
LLM
View more
Python
Machine Learning
LLM
View more
PROJECTS
Automated execution engine for US equities
2026
Automated execution engine for US equities. Detects gap-up and gap-down opens, reconciles missed fills and rebuilds order state without m...
Automated execution engine for US equities. Detects gap-up and gap-down opens, reconciles missed fills and rebuilds order state without manual intervention; survives crashes and reconnects through a PostgreSQL order ledger; uses all-or-none day orders to keep fill pairs symmetric; ships with paper-trade mode and kill-switch controls. Python, Alpaca API, WebSocket and REST.
Python
PostgreSQL
Trading Systems
View more
Python
PostgreSQL
Trading Systems
View more
MEV extraction system on Ethereum and EVM chainsView Project
2026
Off-chain opportunity detection for DEX arbitrage and loan liquidations with on-chain execution through smart contracts, submitted via pr...
Off-chain opportunity detection for DEX arbitrage and loan liquidations with on-chain execution through smart contracts, submitted via private transaction relays to avoid front-running and failed-transaction gas. Months of live research surfaced the real constraints of latency, liquidity and competitive saturation. Python, ethers.js, Solidity, Uniswap v2 and v3.
Python
Solidity
View more
Python
Solidity
View more