Fachrin Aulia Nasution
Location: Berlin, Germany • Work authorization: Permanent Resident • Phone: +491639860973 • Email: [email protected]
Website: https://fachr.in • GitHub: https://github.com/parinpan • LinkedIn: https://linkedin.com/in/fachrinfan
Profile
Senior Product Engineer with 9+ years of experience building systems that serve millions across ecommerce, on-demand mobility, food delivery, and fintech, from ride services in Southeast Asia at Gojek to investment infrastructure in Europe at Upvest. Strongest in owning ambiguous product problems end to end, automating manual workflows, and turning painful systems into reliable, scalable platforms.
Experience
Senior Product Engineer at Upvest • Berlin, Germany
-
- Designed and built a streaming data pipeline that moves entitlement data automatically from BigQuery into the processing engine. Before this, Ops teams had to manually download eligibility and entitlement records from Looker and re-upload them to the processing dashboard. At peak seasons, tens to hundreds of corporate action events per day meant hours of manual work per cycle. The pipeline uses gRPC streaming with Postgres COPY FROM to transfer 300K+ records in seconds, eliminating the manual step entirely. Tech: Golang, BigQuery, gRPC, PostgreSQL
- Designed and built the corporate action instructions system, allowing shareholders to submit their own election choices for voluntary corporate actions (e.g. cash vs. securities for dividends, tender offers) instead of relying on whatever Ops set as static defaults. I also authored the API contracts for clients to place instructions and receive updates via webhooks. This unlocked self-service election capabilities at scale, making Upvest's product offering enterprise-ready for high-AUM clients such as DKB (Deutsche Kreditbank). Tech: Golang, Kafka, PostgreSQL, gRPC
- Designed and built the corporate actions system end to end, from information management through to the processing engine. The information management layer ingests announcements from multiple custodians such as Vestima/Clearstream and BNP Paribas, plus local data providers like WMDaten, then validates, reconciles, and consolidates them into a single golden record that the processing engine depends on for accuracy. Also authored the Public API webhook events contract that clients use for corporate action notification integration. Tech: Golang, Kafka, PostgreSQL, BigQuery, gRPC
- Designed the multi-tenancy architecture for the tax exemption system and authored the Public API contract that enterprise clients integrate against. This gave clients like N26 and Revolut a self-service way to manage tax exemption information for their entire user base, replacing what was previously a manual, Ops-driven process. Tech: Golang, Kafka, PostgreSQL, SFTP, gRPC
- Introduced a Domain-Driven Design (DDD) code architecture for the corporate action service repository, significantly enhancing project execution speed and productivity, and setting a benchmark for effective code architecture within the organisation. Other teams later adopted the same pattern for their own services. Tech: Golang
- Designed and implemented Swift MT564, MT567, MT565, and MT568 message parsers, ensuring full compliance with Swift ISO 20022 standards. This enabled the team to extract and map corporate action notification and instruction details into the domain model, significantly improving data handling and processing efficiency. The parsers feed structured data directly into both the corporate actions pipeline and the instructions management system. Tech: Golang, Swift ISO 20022, Protobuf
- Interviewed software engineering candidates and contributed to the question bank and evaluation criteria the team used to assess them, giving interviewers a consistent set of questions and a shared scoring standard for evaluating candidates fairly.
Senior Software Engineer at Choco • Berlin, Germany
-
- Resolved a critical loading performance issue on product view pages with up to 2k products. Identified and eliminated bottlenecks in GraphQL queries with numerous resolvers, slashing wait times and enhancing user experience and UI metrics. Rewrote the worst-offending queries and collapsed redundant resolver calls, bringing page load time down to a range where the product was actually usable for merchants. Tech: AWS Lambda, GraphQL, DynamoDB
- Designed and executed a high-performance data migration and synchronisation pipeline, seamlessly transitioning data from legacy to performant data sources while ensuring uninterrupted operations. Used DynamoDB Streams and Kinesis to keep old and new data stores in sync during the cutover, so no reads or writes were dropped during the migration window. Tech: DynamoDB Streams, Amazon Kinesis
Software Engineer II at Delivery Hero • Berlin, Germany
-
- Introduced a better mechanism to send metrics to Datadog across the billing domain by writing a proper HTTP and SQS middleware in Golang, allowing teams to improve their monitoring and alerting standards and help them obtain accurate north-star metrics for p50 to p99 distribution of incoming and outgoing HTTP request latency, consumer processing and delay time, and other key processing metrics. The middleware was adopted as the standard across the billing domain, giving every team a consistent way to track latency and set alerts without custom instrumentation. Tech: Golang, Datadog, SQS
Lead Product Engineer at Gojek • Jakarta, Indonesia
-
- Led the launch of new service types such as GoCar Protect, GoRide EV, and GoCar Premium across multiple Southeast Asian markets, driving the initial technical analysis and cross-functional execution. It greatly helped Gojek in gaining a large number of active users and becoming the market's top-of-mind during the pandemic. Tech: Golang, Kafka, Redis, PostgreSQL
- Led the technical analysis and development of Gojek Singapore's Cancellation Fee improvement. It resolved systemic issues with order cancellation caused by incorrect ETA calculations, chained bookings, and driver-induced cancellation. Complaint tickets were reduced by 3%, which also reduced the volume of manual refund reviews the support team had to handle. Tech: Golang, Python, Redis, Kafka, BigQuery, TensorFlow
- Led the development of Driver Karma V2, a feature for communicating information about the driver partner's achievements, feedback scores, and vehicle hygiene to customers. Riders feel safer taking a trip knowing more about the driver, which contributed to stronger customer trust and an improved ride experience on the platform. Tech: Golang, Kafka, Clojure, PostgreSQL, HAProxy
- Researched and developed a machine learning model using BERT to predict the chat intent between drivers and riders during active bookings. It reduced cancellation disputes raised by unfair driver-induced cancellation events, and saved the company ~75K SGD within the first 3 months of the model being shipped. Tech: Golang, Python, Redis, Kafka, BigQuery, TensorFlow
- Initiated and developed an automated testing framework to verify published Kafka messages across event-driven workflows. The framework was integrated into the CI/CD pipeline and subsequently adopted across Gojek product groups like Transport, GoFood, and Logistics, giving those teams an automated way to catch message contract regressions before deployment. Tech: Java, Kafka, Gitlab
- Developed a revamped Tipping Widget for GoCar and GoRide that addressed a discoverability problem where over 90% of orders received no tips. The revised experience introduced on-trip tipping, contextual reminders, emotion-linked tip selectors, and driver humanisation elements. It increased the average tip amount by 100%, and increased the tip frequency by 13% within the first few weeks of launch. Tech: Golang, Redis, Kafka, PostgreSQL, RabbitMQ, HAProxy
- Ran numerous engineering initiatives across the transport platform. For example, reduced Kafka consumer processing time on service deployment from 1 hour to only 5 minutes. The consumer processes roughly 10K messages per minute, so the 55-minute saving was significant at scale. Tech: Clojure, Kafka
Software Engineer L3 at Tokopedia • Jakarta, Indonesia
-
- Worked on SuperInteger, a new integer data type for Tokopedia's Go GraphQL library. Previously, the library only supported 32-bit integers; SuperInteger supports 53-bit integers and was widely adopted across the organisation. Tech: Golang, GraphQL
- Created a database migration layer switching from PostgreSQL to CockroachDB without causing down time. Using CockroachDB helped the team to scale up the database horizontally (sharding). Tech: PostgreSQL, CockroachDB, Redis
- Developed a regression test framework to ensure that recent software changes do not negatively impact the existing flow by running a microservice app in an isolated Docker environment. Tech: Golang, Docker, gRPC, Protobuf, WireMock
- Sped up the shopping cart microservice processing time by 200-400 ms through concurrency. Tech: Golang
Education
Bachelor of Computer Science, University of Sumatera Utara • Medan, Indonesia
-
- Built a low-latency Single Sign-On system that unified authentication across the university's services, serving thousands of employees, students, and alumni (https://akun.usu.ac.id).
- Served as a teaching assistant for Programming Fundamentals, Mobile Programming, Object-Oriented Programming, Data Structures and Algorithms, Internet Programming, and Semantic Web.
Skills
- Programming Languages
- Go, Typescript, Python, Clojure, Java
- Databases
- PostgreSQL, MySQL, DynamoDB, Redis
- Message Queues
- Kafka, SQS, RabbitMQ, NSQ
- Protocols
- HTTP, gRPC, Websocket, WebRTC
- Containerization
- Kubernetes, Docker, Amazon ECS
- Cloud Services
- Google Cloud, AWS
- APIs, Load Balancers, Reverse Proxies
- Kong, HAProxy, Nginx, GraphQL
- Tools
- Git, Terraform, Github Actions, Prometheus, Grafana, Datadog
- AI Coding Tools
- OpenCode, Claude Code, Codex, Omniroute