Kalpit Shukla — Senior Java Engineer

Kalpit Shukla

16 Years · Citi · MarketAxess

Senior Java contractor. Electronic trading depth. No ramp-up required.

✉shukla.kalpit@gmail.com ✆07576 699 940 in LinkedIn

Core Stack

Java 21 FIX 4.4/5.0 Spring Boot Kafka kdb+/q React Angular Oracle Redis Docker LangChain4j AWS AI & ML Python(scripting, data analysis, ML)

Domain

Fixed Income RFQ Equities FX Algo Trading Low Latency Payments Post Trade MiFID II

⬡ Kaltech

⬡ MarketAxess

⬡ Citibank

⬡ Earlier

Kaltech Software Ltd

Dec 2025 – Present

Independent Contractor — Proprietary Trading Technology

Equity Signal Engine

Java 21 · Spring Boot 3 · Virtual Threads · LangChain4j · ta4j · React · REST APIs

Built a real-time signal computation engine scanning the S&P 500 universe — multi-provider market data integration with configurable abstracted REST client layer.
Technical indicator computation (RSI, MACD via ta4j) for pre-trade signal generation across 500+ instruments simultaneously.
Virtual-thread executor service managing ~500 concurrent instrument fetches with rate-limit-aware scheduling using scheduleAtFixedRate batching — no CountDownLatch overhead, deterministic throughput under provider constraints.
Sub-100ms signal generation latency across 500 concurrent instruments.
LangChain4j LLM integration for AI-powered signal rationale generation using Claude API via AWS Bedrock.
React frontend with multi-dimensional scoring heatmap (T/F/C/B/G/E/S/M dimensions) powered by Nivo charts.

MarketAxess

Mar 2020 – Nov 2025 · Contract then Permanent (Jul 2022)

Senior Software Engineer — Front Office Fixed Income Bond Trading

Bondlink — Fixed Income RFQ Trading Platform

Java 11-21 · Spring · FIX Protocol · IBM MQ · JMS · Kafka · Oracle · Shell Script

Delivered key enhancements for UK and Europe Fixed Income market segments — UK Gilts, European Credit, EM bonds.
Contributed across multiple trading flows: RFQ Open Trading, Disclosed, List, Crossing, AXE, Portfolio, and IOI — price/spread protocols for One Step, Spot, and Phone Trade execution paths.
Covered full trade lifecycle from RFQ execution through to post-trade booking confirmation and settlement instruction generation via FIX — T+1/T+2 cycles.
Modernisation of legacy Bondlink components to JDK 21.
Owned production release cycle for Bondlink UK/EU segments — release coordination, deployment verification, post-release monitoring.
Used Claude via AWS Bedrock to document FIX workflows — test coverage from under 10% to 45%, reducing Checkmarx security vulnerabilities.

REMS — Real-Time Exception Management Surveillance

React · Java 21 · Spring Boot · REST APIs · Redis · Docker · Kubernetes · AWS

Designed REMS post-trade exception management platform — trade break identification, failed settlement resolution, allocation updates.
ORK-to-ARM record conversion and FCA submission covering MiFID II RTS22 transaction reporting, RTS23 reference data, RTS24 order record keeping.
React/AG Grid dashboard with server-side rendering, Spring Boot REST with Spring Cloud Circuit Breaker, Spring Data on AWS RDS.

Citibank

Feb 2011 – Jun 2018 · 7 Years

Senior Core Java Developer (AVP) — Front Office Equities & FX

CITI SMART — Algorithmic Trading Engine

Java 8 · Spring 3.5 · LMAX Disruptor · FIX 4.4 · TIBCO EMS · kdb+ · Bloomberg · Reuters · Unix

Engineered low-latency performance improvements — replaced synchronized blocks with lock-free data structures and java.util.concurrent primitives (ConcurrentHashMap, Atomic classes, tryLock with timeout) eliminating lock contention.
Consumed LMAX Disruptor v3 ring buffer within the CFore→SMART order pipeline, processing inbound FIX orders via BusySpinWaitStrategy on a dedicated isolated CPU core.
Optimised critical path execution by identifying and removing contention hotspots in order routing and fill processing — reduced thread-blocking on shared state.
Profiled JVM heap allocation and GC behaviour using Java Flight Recorder and async-profiler — applied object pooling to reduce GC pressure on critical execution path.
Engine consumed orders from CFORE and routed to market using Bloomberg and Reuters real-time market data via CitiMatch internal dark pool.

Total Touch / Liquifi / CitiMatch — Dark Pool

Java · FIX 4.4 · TIBCO EMS

Delivered IOI features for Total Touch — the in-house liquidity provider and internal crossing engine for Citi Equity business.
CitiMatch served as the integrated dark pool execution venue and test ecosystem for CITI SMART.

FX Options Pricing

Java · kdb+/q · Bloomberg · Reuters · TIBCO EMS

Implemented FX options pricing data services — extended data models with options-specific attributes (strike price, expiry date, option type, Greeks: Delta, Gamma, Theta).
Integrated real-time spot rates from Bloomberg and Reuters, published enriched pricing data to front-office trading systems via TIBCO EMS.
Updated kdb+ schema for FX options data with efficient partitioning and indexing for real-time analysis.

Earlier Contracts & Roles

Jun 2007 – Feb 2011 / Jul 2018 – Mar 2020

Java Developer / Lead / Application Developer

JAJA Finance — Microservices Migration

Java · Spring Boot · Kafka · AWS · ISO 8583 · ISO 20022

Led migration from monolithic to event-driven microservices for a challenger bank payments platform.
Implemented ISO 8583, ISO 20022 message translation layer and real-time payment processing services.

TP ICAP — Execution Report Distribution

Java · Spring Boot · Solace · Hazelcast · AWS S3

Two-node distributed execution report aggregation service — deduplication by orderId/tradeId, zip generation, S3 upload, automatic failover to secondary node.

10xBanking / Infosys — FX Rate Processing

Java · Spring Boot · AWS · Docker · Kubernetes · Grafana

Real-time FX rate ISO 20022 message processing services on AWS with zero-downtime Kubernetes deployment.

3i-Infotech — Credit Suisse / ABSA Barclays

Java · Spring · Oracle

Extended KASTLE Factoring platform for Credit Suisse and ABSA/Barclays — onsite engagement at Credit Suisse, Zurich.

◈ The Story

Kaltech Software Ltd

The engine was not a tutorial project. It manages ~500 concurrent instrument fetches using Java 21 virtual threads, with rate-limit-aware scheduling that ensures deterministic throughput without CountDownLatch overhead. The data ingestion layer abstracts multiple market data providers behind a configurable REST client — meaning swapping providers requires zero code changes.

I integrated LangChain4j and the Claude API to generate AI-powered signal rationale — giving each technical indicator signal a human-readable explanation. This was not a demo. It runs, it produces output, and it informed my own investment decisions.

The React frontend renders a multi-dimensional scoring heatmap across Technical, Fundamental, Catalyst, Behavioural, Growth, ESG, Sentiment and Momentum dimensions.

Hello. I'm Kalpit — ask me anything about my experience, the role shown, or my availability.

just now