Personal details

Satwik K. - Remote back-end developer

Satwik K.

Timezone: New Delhi (UTC+5.5)

About

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

Work Experience

Founder & Lead Developer
Zobyt | Jan 2022 - Present
Python
Trading Systems
LLM
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.
Lead Developer - Automated Trading Systems
Hyperion Fintech AG | Dec 2020 - Jun 2024
Python
Algorithmic trading
Trading Systems
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.

Projects

Automated execution engine for US equities
2026
Python
PostgreSQL
Trading Systems
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.
MEV extraction system on Ethereum and EVM chains
2026 | https://www.zobyt.com/work/automated-mev-extraction-system-on-ethereum-and-evm-compatible-blockchains
Python
Solidity
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.