Quantman is a no-code algorithmic trading platform designed for intraday stock traders. It allows them to automate their trading strategies, helping to eliminate emotional decision-making and ensuring more consistent, disciplined trading

• Architected and built a composable DSL for expressing trading strategies, enabling both backtesting and live execution from the same strategy definition. Resulted in 10x faster backtesting and 20x reduction in live compute.
• Designed a low-latency order execution pipeline minimizing tick-to-trade time, reducing slippage and improving fill prices in live intraday trading
• Built a conversational agent that translates natural-language strategy descriptions into the composable DSL, enabling non-technical traders to author and iterate on strategies without writing code.

LeadIQ is a people database that helps sales teams identify and target the right individuals at the right companies to boost product outreach. It also provides tools for crafting personalized emails and content, making the sales process more efficient and successful.

• Led adoption of a lakehouse architecture on Databricks, replacing a manual process of dumping MongoDB data onto EC2 instances for aggregation — eliminating hours of ad-hoc work and enabling self-serve analytics.
• Optimized user-facing Elasticsearch queries, achieving 10x latency reduction across critical search paths.

StackState is a cloud-native monitoring framework and a pioneer in AIOps. It features a custom transactional graph database built on top of HBase, enabling advanced capabilities like microservice topology mapping and root cause analysis. These features provide deep insights and empower the monitoring of large, complex cloud-native applications.

• Architected "StackPacks" — modular, installable configuration bundles that let users onboard new environments in minutes instead of hours.
• Tuned Elasticsearch metric queries for a 3x latency reduction, directly lowering platform-wide resource consumption.
• Designed and built the anomaly and event storage layer, optimizing retrieval for real-time root cause analysis.
• Built a progressive notification system that streams evolving root-cause analysis to Slack, keeping responders in the loop as diagnosis unfolds.

Picnic is an online supermarket in the Netherlands and parts of Germany, known for its strong environmental focus. They were early adopters of electric vehicles for delivery, even before EVs became mainstream. Additionally, they optimize their delivery process by grouping orders into time slots, ensuring efficient and sustainable delivery to customers

• Designed a DSL for expressing and enforcing fair-usage policies on delivery slots, preventing systematic abuse of the scheduling system.

Indix is a product and price database that crawls various e-commerce sites to collect product and pricing data. This information is then used by other e-commerce platforms to competitively price their products or enhance their product coverage. By leveling the playing field, it allows smaller e-commerce businesses to compete with larger ones. Indix also explored multiple innovative business models, offering unique value in the market.

• Designed a crawl scheduler using kinetic heaps to model URL priority as a function of time, ensuring optimal re-crawl frequency and data freshness across millions of product pages.
• Rewrote crawler infrastructure with non-blocking I/O — 4x throughput, half the machines. Open-sourced the underlying library as bubblewrap.
• Built an event-sourced configuration system for the crawler, enabling full audit trails and rollback of crawl settings.
• Built a DSL and web application for authoring product-extraction parsers, significantly reducing time-to-onboard new e-commerce sites.
• Co-built an automated wrapper induction system that generates product parsers from page structure, replacing manual parser authoring for new site onboarding.
• Built an in-memory analytics engine for real-time price aggregations over time-series data.
• Built a cross-store product matching system enabling price comparison and competitive positioning across e-commerce catalogs.

Worked at ThoughtWorks, a premium software consulting firm, on multiple projects primarily using the JVM/Spring tech stack. As a consultant, I was not only responsible for designing solutions but also for maintaining constant communication with stakeholders and clients to ensure alignment, address concerns, and drive project success.

• Built an inventory management web app for a retail chain, targeting legacy handheld devices with severe browser constraints.
• Built a procurement platform for an aircraft manufacturer's technical manuals, serving customers and maintenance teams.
• Built a mobile phone and contract comparison app using progressive enhancement, ensuring functionality without JavaScript.

Intel introduced a new mobile processor called Atom, and I was part of a small team that built devices to showcase its capabilities. These devices helped demonstrate the power and potential of the Atom processor in real-world applications.

• Built proof-of-concept applications on custom Intel Atom small-form-factor devices (smart card reader, fingerprint scanner, thermal printer), working across Linux Kernel and Windows Embedded.