Chicago, Illinois / Applied AI
I’m Jacky, a Senior Data Scientist in Chicago. I build useful AI products—and bring people, evidence, and business priorities together to make them work.
Jinghang Li, Ph.D.AI · Product · Leadership
Chicago, IL · 41.8781° N
Introduction / About
I don’t believe there is one best answer for every situation. I start with the evidence, understand the constraints, and find the solution that fits the people and the problem.
My work spans AI development, product decisions, and collaboration with engineering and product teams across the US, China, Malaysia, and Singapore. I also mentor ML engineers, interview technical candidates, and help teams turn ideas into working prototypes through internal AI hackathons.
Selected work / Pipeline performance
I led an update to an existing request-processing pipeline. The goal was straightforward: make a slow workflow useful at the pace people needed it.
As adoption grew, a single-threaded design became the next bottleneck. A full rewrite carried migration and versioning risks. I found a way to increase concurrency within the existing system, avoiding a disruptive re-architecture.
The upgraded pipeline supports 20+ concurrent requests. Estimated potential throughput grew from hundreds of requests per day to 5,000+ per day—a capacity estimate, not a claim of observed daily usage.
Product judgment / Team leadership
01 / Call-center AI
40%less task-processing time per call
Pilot result
I aligned AI, engineering, frontline feedback, and product requirements across international teams. We moved beyond knowledge search to concise, referenced guidance that representatives could use during a live call.
Inside the cross-team project02 / Computer-use workflows
>40%fewer tasks needing manual handling
Within the covered operations workflow
I turned two documented paths into a map of 10+ workflow branches through hands-on testing and operational walkthroughs, then divided the work across our AI team. We delivered ahead of schedule.
From discovery to deliveryIn the field / Leadership & perspectives
Recognition / Beyond the individual contribution
I began as an AI subject-matter expert, helping hackathon teams turn ambitious ideas into stable proofs of concept. My role expanded into shaping challenge topics with leadership and judging submissions.
The question became bigger than “Can we build it?” I look for a real business problem, a practical delivery path, and a credible benefit beyond the demo.
How my leadership approach evolvedI discuss connecting customer needs to practical AI workflows, and why product understanding matters as much as technical depth.
Featured with Daniel Scott · Written by Taylor Rose
Read the interviewA conversation about continuous learning, reflection, and connecting AI with operations research to support product and business decisions.
Featured interview · Written by Taylor Rose
Read the interviewChapter ISystems that converse, reason, and act
LLM systems built around routing, grounding, evaluation, and human judgment.
Explore projects01 / Production AI
My work connects orchestration, knowledge retrieval, and evaluation with product needs. A useful agent should help a person take the next step, with a clear path to review when it is uncertain. My current work includes multi-agent workflows in a platform supporting 700K+ monthly AI requests— the platform’s operating scale, not an individual performance result.
Use the smallest capable path and make delegation visible.
Bring policy, knowledge, and operational context into the response.
Test usefulness, risk, and consistency—not only model fluency.
Expose confidence, reversibility, and a clear review path.
Chapter IIFrom a large network to the next best match
Scoring systems that combine fit, constraints, availability, and real operating context.
Explore projects02 / Decision systems
My work joins candidate quality, operational eligibility, and changing network conditions into a ranked action that users can understand and improve.
Eligibility, availability, and operational needs belong in the decision—not outside the model.
A better score is useful only if it leads to a better next action.
Chapter IIILearning when the world changes
Online meta-learning and feedback loops for systems that cannot assume tomorrow looks like yesterday.
View the research03 / Applied ML research
My doctoral work explored continuous model adaptation using online meta-learning—research that still shapes how I think about monitoring, evaluation, and feedback in production AI.
Watch the relationship between current data and learned assumptions.
Update what needs to move without forgetting what still works.
Evaluate the system across time, not just at a frozen benchmark.
Chapter IVResearch depth, production pace
Research, production delivery, and the cross-functional work that turns technical progress into business value.
Open the timelineAI product initiatives, global collaboration, mentoring, and hands-on technical delivery.
04 / Experience
Leading AI product initiatives across a property-services platform, collaborating with international engineering and product teams, mentoring ML engineers, contributing to technical hiring, and helping lead internal AI hackathons.
Delivered scoring and recommendation work across approximately 100K locations and 10K market/trade combinations, alongside ML pipelines and cloud data workflows.
Supported data-centric applications and Data as a Service capabilities using Python, SQL, and applied ML patterns.
Researched online meta-learning, deep-learning optimization, and continuous model adaptation for smart-grid applications. Ph.D. awarded in December 2021; research continued through January 2022.
Where practice meets discoveryNext / Research
Chapter VIdeas built to travel
Academic work in adaptive intelligence, carried forward into practical AI design.
Read the record05 / Selected research
Peer-reviewed work, a doctoral thesis, and a continuing interest in systems that learn under real-world change.
Adaptive prediction under shifting training and real-time data patterns.
View IEEE record ↗ Doctoral thesis · UICOnline meta-optimization, deep neural networks, and real-time change.
Read the thesis ↗ Research profileA question, an idea, or a complex system?
I’m glad to compare notes on AI products, practical data problems, and building teams that turn good ideas into useful work.
[email protected] ↗